E9sht I C C I W D S

Internal Corrosion Control in Water Distribution Systems
Improve water quality at customer taps
W ater professionals are very familiar with the problems caused by internal pipe corrosion: water quality
Internal Corrosion
degradation, customer complaints, reduced pipe capacity, and home plumbing failures. Surprisingly,
however, there is little practical guidance available to public water systems regarding the design, implementation,
and maintenance of an ongoing internal corrosion control program.
Control in Water
These issues prompted the publication of this new AWWA manual of water supply practices. With this practical
manual, you will know how to control internal corrosion in metal pipes and plumbing pipes, ensure compliance
with the USEPA’s Lead and Copper Rule, and provide the best water quality to your customers.
Distribution Systems
M58 covers everything you need to know about internal corrosion
• Causes of internal corrosion
• Pipe problems and aesthetic and health issues caused by internal pipe corrosion
• Planning and implementing a monitoring and control program
• Advantages, disadvantages, and comparative costs of corrosion control chemicals
• Performing bench tests and pilot tests
M58
• Maintaining excellent water quality between the treatment plant and customer taps
M58
Manual of Water Supply Practices
First Edition
Advocacy
AWWA is the authoritative resource for knowledge, information, and advocacy to improve the quality and
Communications
supply of water in North America and beyond. AWWA is the largest organization of water professionals in Conferences
the world, advancing public health, safety, and welfare by uniting the efforts of the full spectrum of the Education and Training
water community. Through our collective strength, we become better stewards of water for the greatest Science and Technology
good of people and the environment. Sections
1P-5C-30058-12/10-SB The Authoritative Resource on Safe Water ®
M58 cover1.indd 1 11/17/2010 4:38:11 PM

Internal Corrosion Control
in Water Distribution Systems
AWWA MANUAL M58

First Edition
M58 book.indb 1 11/17/2010 4:24:29 PM

MANUAL OF WATER SUPPLY PRACTICES — M58, First Edition
Internal Corrosion Control in Water Distribution Systems
Copyright © 2011 American Water Works Association
All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopy, recording, or any information or retrieval system, except in the
form of brief excerpts or quotations for review purposes, without the written permission of the publisher.
Disclaimer
The authors, contributors, editors, and publisher do not assume responsibility for the validity of the content
or any consequences of their use. In no event will AWWA be liable for direct, indirect, special, incidental, or
consequential damages arising out of the use of information presented in this book. In particular, AWWA will
not be responsible for any costs, including, but not limited to, those incurred as a result of lost revenue. In no
event shall AWWA’s liability exceed the amount paid for the purchase of this book.
Project Manager: Melissa Valentine

Editor: Deborah Lynes
Production Editor: Cheryl Armstrong
Manuals Coordinator: Molly Beach
Library of Congress Cataloging-in-Publication Data.
Hill, Christopher P.
Internal corrosion impacts in drinking water distribution systems / by Christopher P. Hill, Abigail F.
Cantor. -- 1st ed.
p. cm. -- (AWWA manual ; M58)
Includes bibliographical references and index.
ISBN 978-1-58321-790-0 (alk. paper)
1. Water-pipes--Corrosion. 2. Drinking water--Contamination. 3. Water--Distribution. 4. Corrosion and
anti-corrosives--Testing. I. Cantor, Abigail F. II. Title.
TD491.H64 2010
628.1’44--dc22
2010016956
Printed in the United States of America

American Water Works Association
6666 West Quincy Ave.
Denver, CO 80235 Printed on recycled paper
M58 book.indb 2 11/17/2010 4:24:29 PM

Contents
List of Figures, v
List of Tables, ix
Preface, xi
Acknowledgments, xiii
Chapter 1 Overview of Internal Corrosion Impacts

in Drinking Water Distribution Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Introduction, 1
Purpose of This Manual, 2
Customer and Infrastructure Impacts of Internal Corrosion, 2
Regulatory Impacts, 6
Additional Reading, 9
References, 10
Chapter 2 Fundamentals of Internal Corrosion and Metal Release . . . . . . . . . 13

Introduction, 13
Mechanisms of Metal Release by Uniform Corrosion, 13
Mechanisms of Metal Release by Nonuniform Corrosion, 23
Other Mechanisms of Metal Release, 27
Summary, 28
References, 28
Chapter 3 Water Quality Monitoring and Assessment of Internal

Corrosion and Increased Metals Concentrations . . . . . . . . . . . . . . . . . . . . . . 31
Introduction, 31
Water Quality Considerations, 32
Developing a Water Quality Monitoring Program, 49
Assessing the Cause of Internal Corrosion and Metals Release, 53
Summary, 55
References, 57
Chapter 4 Corrosion Control Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Introduction, 61
Distribution System Design Considerations, 62
Chemical Treatment, 74
Conclusion, 100
References, 101
iii
M58 book.indb 3 11/17/2010 4:24:29 PM

Chapter 5 Implementing Corrosion Control Treatment. . . . . . . . . . . . . . . . . . 103
Introduction, 103
Corrosion Indexes, 105
Bench Testing, 105
Examination of Pipe Scales, 109
Pipe Loops, 110
Coupon Studies, 115
PRS Monitoring Stations, 117
Electrochemistry Loops, 120
Premise Plumbing Profiles, 126
Reservoir Profiles, 127
Summary, 128
References, 129
Chapter 6 Conducting Pilot Studies and Monitoring

Effectiveness of Corrosion Control Treatment. . . . . . . . . . . . . . . . . . . . . . . 131
Introduction, 131
Conducting a Distribution System Pilot Study, 132
Monitoring the Effect of Corrosion Control Treatment, 140
Summary, 148
References, 148
Appendix A Achieving pH Stability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Appendix B A Case Study: Causes of and Actions Taken

to Control Lead Release in the D.C. Distribution System. . . . . . . . . . . . . . . 157
Appendix C North American Corrosion Control Needs and Strategies . . . . . 173
Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
iv
M58 book.indb 4 11/17/2010 4:24:29 PM

Figures
1-1 Red water sample from a US distribution system, 3

1-2 Number of large US water systems exceeding the lead action level, 7
1-3 Comparison of 90th percentile LCR monitoring data
for 166 large public water systems, 7
2-1 Typical battery, 15

2-2 Typical water and metal pipe “battery”, 15
2-3 Barrier between water and metal pipe, 16
2-4 Characteristic horseshoe-shaped pit of erosion corrosion, 25
2-5 Example of microbially influenced pitting of copper pipe, 26
2-6 Another example of microbially influenced pitting of copper pipe, 26
3-1 Variation of buffer intensity with pH, 34

3-2 Relationship between alkalinity and DIC for various pH levels, 35
3-3 Effect of DIC on buffer intensity, 36
3-4 Iron Pourbaix diagram for a carbonate-containing water at 25ºC and I = 0, 40
3-5 Impact of chloride-to-sulfate ratio on lead corrosion, 42
3-6 Precipitated iron as a protective barrier for microorganisms, 48
4-1 Asbestos–cement pipe, 65

4-2 Bronze fitting, 65
4-3 Brass faucet, 65
4-4 Copper pipe, 65
4-5 Galvanized pipe, 66
4-6 Unlined cast-iron main, 67
4-7 Harvested lead service lines, 67
4-8 Plastic water supply pipe, 67
4-9 Branched (A) and grid/loop (B) designs, 68
4-10 Application of cement–mortar lining, 71
4-11 Example of pipe before and after cleaning and lining, 71
4-12 Steel pipe with epoxy coating, 72
4-13 Typical liquid chemical feed system, 74
4-14 Typical dry chemical feed system, 75
4-15 Iron and manganese solubility, 82
4-16 Slurry slaker, 83
4-17 Paste slaker, 84
4-18 Ball mill lime slaker, 84
4-19 Examples of chemical inhibitors, 86
4-20 Three polyphosphate structures, 87
4-21 Type of phosphate inhibitor used by water utilities, 90
4-22 Reasons for using phosphate inhibitors, 90
4-23 Orthophosphate feed system, 90
5-1 Diagram of pH adjustment chemical injected into line before water from
hydropneumatic tank is exposed to atmospheric pressure, 106
5-2 Apparatus setup for filtering calcium carbonate and other
precipitates such as iron, manganese, and phosphate, 107
5-3 Pipes with lead-dioxide scales cut horizontally prior to XRD, 109
M58 book.indb 5 11/17/2010 4:24:29 PM

5-4 Lead-speciated XRD patterns, 110
5-5 Example of a capped pipe ready for transportation, 110
5-6 Circulating-loop schematic, 111
5-7 Circulating loop with lead service lines, 112
5-8 Lead release during pipe-loop conditioning from pipe
containing lead-oxide scales, 112
5-9 Metals release over time in a stagnated lead pipe loop, 113
5-10 Flow-through system, view 1, 114
5-11 Flow-through system, view 2, 115
5-12 Data from a pipe loop that displays changes in lead (IV) release as the
loop was alternated with chlorine and chloramine over time, 115
5-13 Typical metal coupons, 117
5-14 Typical coupon study apparatus, 117
5-15 Drawing of a PRS monitoring station, 119
5-16 Stacks of metal plates exposed to water replace pipe
loops in the PRS monitoring stations, 119
5-17 Electrochemistry circulation pipe loop setup, 124
5-18 Schematic of electrochemical polarization cell used
for corrosion rate measurements, 124
5-19 Typical polarization cell design, 125
5-20 Corrosion rates from lead pipe with lead (IV) scale, 125
5-21 Example of a lead profile at a residential home with a lead service line, 127
5-22 Example of a lead profile at a residential home with galvanized
internal plumbing and lead service line, 128
6-1 Example of maintaining chemical (pH) stability at the entry point, 133

6-2 Percent positive total coliforms in DC Water in relation
to the addition of orthophosphate, 134
6-3 Precipitated phosphate, 135
6-4 Depth samplers used to collect grab samples from reservoirs, 135
6-5 HPC results taken from a routine inside monitoring tap and
a hydrant located next to the property, 136
6-6 Minimum alkalinity required to maintain a buffer intensity greater
than or equal to 0.10 meq/L as a function of pH at 20°C, 141
6-7 Lead profile water hammer graph taken from residential homes in the
District of Columbia from December 2003 to July 2005, 142
6-8 Potential-pH diagram for 1 mg/L free chlorine, showing the speciation
of the chlorine system, the high ORP necessary for free chlorine
stability, and the relationship to the water stability boundary, 142
6-9 Flowchart example of action taken when a regulated
parameter is exceeded causing an excursion, 146
6-10 SOP flowchart for exceeding target levels, 147
A-1 Example of a dual-speed pump calibration curve, 150

A-2 WQP operations chart for pH/alkalinity adjustment, 152
A-3 WQP operations chart: soda ash and target pH of 7.8, 152
A-4 Filling in WQP chart with sample data, 154
A-5 Working WQP operations chart, 154
vi
M58 book.indb 6 11/17/2010 4:24:29 PM

B-1 The Washington Aqueduct service area, 158
B-2 History of LCR compliance in Washington, D.C., 160
B-3 Peak dissolved lead levels in homes during lead profiles, 161
B-4 EMF-pH diagram for Pb-H2O-CO2 system, 162
B-5 DC Water circulation loop testing results, 163
B-6 Location of partial system test, 165
B-7 Photos of temporary phosphoric acid feed equipment, 165
B-8 Fact Sheet: New Treatment to Address Lead in Water, 166
B-9 Excerpts from DC Water’s test sampling and response plan, 167
C-1 Geographic location of US participants by USEPA region, 174

C-2 Percentage of CWSs in SDWISFED versus percentage of
survey respondents by USEPA region, 175
C-3 Geographic location of participants by Canadian province/territory, 175
C-4 Number of systems by population category, 176
C-5 Range of source water pH, 176
C-6 Range of source water alkalinity, 177
C-7 Range of source water DIC, 177
C-8 Number of systems practicing and not practicing
corrosion control for Source #1, 178
C-9 Corrosion control treatment used for Source #1, 180
C-10 Percentile distribution plot of US 90th percentile lead levels, 180
C-11 Percentile distribution plot of US 90th percentile copper levels, 181
C-12 90th percentile lead levels as a function of pH, 182
C-13 90th percentile lead levels for systems using phosphates, 182
C-14 90th percentile lead levels as a function of type of phosphate used, 183
C-15 Comparison of 90th percentile lead levels for systems
practicing pH adjustment versus phosphates, 183
C-16 90th percentile copper levels as a function of pH, 184
C-17 90th percentile copper levels for systems using phosphates, 184
C-18 Comparison of 90th percentile copper levels for systems
practicing pH adjustment versus phosphates, 185
C-19 Percentile distribution of average lead and copper
levels for Canadian respondents, 187
vii
M58 book.indb 7 11/17/2010 4:24:29 PM

This page intentionally blank.
M58 book.indb 8 11/17/2010 4:24:30 PM

Tables
1-1 Eight steps to implementing an effective corrosion control program, 1

1-2 Summary of potential copper corrosion issues, 5
2-1 Chapter 2 key points, 14

3-2 Impact of coagulant changes on lead corrosion, 42
3-3 Suggested water quality monitoring parameters, 50
3-4 Assessment of common corrosion-related water quality problems, 54

4-2 Corrosion properties of materials frequently used in water distribution systems, 63
4-3 Galvanic couples in the water industry that are dangerous, 70
4-4 Pipe wall linings, 72
4-5 Common corrosion control chemicals, 73
4-6 Treatment chemical water quality and corrosion control aspects, 75
4-7 Chemical operational aspects, 79
4-8 Comparative cost of treatment chemicals, 93
4-9 Relative cost information, 93
4-10 Potential impact of treatment changes on corrosion control, 96
4-11 Potential secondary impacts, 98
4-12 Corrosion control technique selection criteria, 100

5-2 Summary of coupon protocols used for distribution
system corrosion measures, 118
5-3 Summary of EC corrosion assessment methodologies, 121
5-4 EC corrosion data analysis software, 122

6-2 Example of distribution target levels for water quality parameters
routine monitoring during pilot testing, 139
6-3 Example of target and excursion levels for OCCTWQPs, 145
A-1 Summary of pH stability requirements, 150

A-2 Alkalinity increment as dictated by average alkalinity baseline value, 153
B-1 Key characteristics of the D.C. distribution system, 159

B-2 DC Water distribution system target levels, 171
C-1 Breakdown of survey responses, 174

C-2 Typical source water quality parameters—utilities
considering corrosion control, 178
C-3 Purpose for corrosion control at Canadian utilities
with a corrosion control program, 179
C-4 Summary of sample approaches by province/territory, 186
ix
M58 book.indb 9 11/17/2010 4:24:30 PM

M58 book.indb 10 11/17/2010 4:24:30 PM

Preface
Volunteers from the Distribution System Water Quality Committee of the American
Water Works Association (AWWA) have prepared this manual of practice. The need for
a manual on corrosion assessment, monitoring, and control results from the increased
focus on corrosion-related water quality and infrastructure impacts in drinking water dis-
tribution systems. Considerable literature is available regarding the factors that influence
corrosion in the distribution system. However, little practical guidance is available to pub-
lic water systems regarding the design, implementation, and maintenance of an effective
corrosion control program. This manual seeks to provide that practical guidance.
This manual helps readers understand the factors that influence corrosion, assess
corrosion-related impacts, and develop a strategy to implement and maintain effective
corrosion control in the water distribution system.
The manual is organized in three main parts. Chapters 1 through 3 help the reader
develop an understanding of the factors that influence corrosion and determine potential
causes of corrosion in the distribution system. Chapters 4 and 5 present corrosion control
alternatives and outline development of an effective corrosion control strategy. Chapter 6
discusses monitoring and optimization to maintain effective corrosion control treatment.
Appendix A provides an example of how to achieve and maintain stable pH in the distribu-
tion system for the purposes of maintaining effective corrosion control. Appendix B pro-
vides a very thorough case study of one utility’s approach to identifying the causes of and
successfully dealing with corrosion-related water quality problems. Each chapter may be
read stand-alone. Therefore, there is some limited repetition among the chapters to furnish
necessary background of important concepts; references to other chapters are provided.
Appendix C is a summary of the 2008 AWWA-DSWQC Corrosion Survey sponsored by the
American Water Works Association (AWWA) Water Quality & Technology Division’s Distri-
bution System Water Quality Committee (DSWQC) and funded by the AWWA Technical and
Education Council, conducted in the fall of 2008. This web-based survey summarized cor-
rosion control objectives, practices, effectiveness of practices (US utilities), and corrosion
control needs (primarily for Canadian utilities) for over 150 utilities in North America.
As previously mentioned, this manual is intended to be a practical guide to imple-
mentation of an effective corrosion control program. As a result, it provides only an over-
view of available research and literature in many areas. To assist the reader in identifying
additional research or literature that may be of interest, many chapters include additional
suggested readings that might be of interest and that will provide more detail regarding a
number of the main topics contained in that chapter.
The materials included herein provide a compendium of the state-of-the-art
knowledge as of the writing of this manual. The renewed focus on corrosion in the
distribution system has resulted in much new research in this area. As a result, more
data and a better understanding of some key corrosion concepts are now being developed,
including: the limited applicability of calcium carbonate saturation indices as predictors
of corrosion; an increased focus on the prevalence of microbially influenced corrosion;
the role of coagulant change, chloride, and sulfate in corrosion; the importance of
oxidation-reduction potential; and better tools for evaluating and managing corrosion
control effectiveness. As a result, this manual will likely be updated in the future to reflect
these advances and better understanding of corrosion and corrosion control treatment in
drinking water distribution systems.
Editor’s note: Throughout this manual, references are made to the Awwa Research
Foundation (AwwaRF), the original name of the foundation. As of 2009, the foundation
changed its name to the Water Research Foundation. Any publication prior to 2009 will
reflect the foundation’s original name.
xi
M58 book.indb 11 11/17/2010 4:24:30 PM

M58 book.indb 12 11/17/2010 4:24:30 PM

Acknowledgments
The first edition of M58 was written through the persistent, dedicated work of the
following authors:
A. Cantor, Process Research Solutions LLC, Madison, Wisc.

R. Giani, DC Water, Washington, D.C.
C. Hill, Malcolm Pirnie Inc., Tampa, Fla.
S. Reiber, HDR Inc., Bellevue, Wash.
E. Turner, City of Dallas, Dallas, Texas
The authors would like to acknowledge the support of the following organizations in
preparing this manual:
American Water Works Association, Denver, Colo.

Dallas Water Utilities, Dallas, Texas
District of Columbia Water, Washington, D.C.
Malcolm Pirnie Inc., Tampa, Fla.
Process Research Solutions, LLC, Madison, Wisc.
The authors would also like to acknowledge the following individuals who provided
editorial and technical comments:
Y. K. Cohen, Southwest Water Company, Covina, Calif.

A. Wilczak, San Francisco Public Utilities Commission, San Francisco, Calif.
The following individuals provided peer review of this manual. Their knowledge and
efforts are gratefully appreciated:
G. Boyd – HDR Engineering, Bellevue, Wash.

G. Burlingame, Philadelphia Water Department, Philadelphia, Pa.
I. Douglas, City of Ottawa, Ottawa, Ontario, Canada
J. Dyksen – United Water, Oradell, N.J.
K. Kunihiro – Orlando Utilities Commission, Orlando, Fla.
F. Lemieux – Health Canada, Ottawa, Ontario, Canada
D. Lytle, USEPA, Cincinnati, Ohio
D. Metz, Greater Cincinnati Waterworks, Cincinnati, Ohio
K. Moriarty – Bangor Water District, Bangor, Maine
T. Pajor – City of Wichita Water and Sewer Department, Wichita, Kan.
S. Reiber, HDR Engineering, Bellevue, Wash.
M. Schock, USEPA, Cincinnati, Ohio
M. Smith, Philadelphia Water Department, Philadelphia, Pa.
This manual was approved by the AWWA Distribution System Water Quality Commit-
tee. Members of the committee at the time of approval of this 1st edition were as follows:
G. Burlingame, Philadelphia Water Department, Philadelphia, Pa.

J. Dyksen, United Water, Oradell, N.J.
C.Hill, Malcolm Pirnie, Inc., Tampa, Fla.
K. Hodsden, MWH Americas, Bridport, Vt.
R. Giani, DC Water, Washington, D.C.
xiii
M58 book.indb 13 11/17/2010 4:24:30 PM

A. Job, Grand Forks Water Treatment, Grand Forks, N.D.
S. Liu, Camp, Dresser, & McKee Inc., Maitland, Fla.
F. Mahmood, Malcolm Pirnie Inc. Dallas, Texas
Q. Muylwyk, CH2M HILL, Toronto, Ontario, Canada
M.L. Nguyen, Nevada Division of Envrironmental Protection, Carson City, Nev.
M. Roberts, Hazen & Sawyer, Greensboro, N.C.
J. Routt, Jan Routt & Associates LLC, Lexington, Ky.
C. Schreppel, Mohawk Valley Water Authority, Utica, N.Y.
J. Skadsen, Camp, Dresser & McKee Inc., Ann Arbor, Mich.
V. Speight, Latis Associates, Arlington, Va.
A. Spiesman, U.S. Army Corps of Engineers Washington Aqueduct, Washington, D.C.
M. Sushynski, Sensicore, Fowlerville, Mich.
K. Thompson, CH2M HILL, Englewood, Colo.
E. Turner, City of Dallas, Dallas, Texas
R. Vaidya, Camp, Dresser & McKee Inc., Tampa, Fla.
S.L. Wagner, Newark Water Department, Newark, Ohio
L. Weinrich, American Water, Delran, N.J.
xiv
M58 book.indb 14 11/17/2010 4:24:30 PM

AWWA MANUAL M58
Chapter 1
Overview of Internal
Corrosion Impacts
in Drinking Water
Distribution Systems
Christopher P. Hill
Malcolm Pirnie Inc.
Introduction_ _________________________________________________________
Prior to development of a corrosion control program, it is important to first gain an under-
standing of the factors that influence internal corrosion and metal release in drinking
water distribution systems. Afterward, implementation of an effective corrosion control
program can be accomplished in eight steps (Table 1-1).
Table 1-1 Eight steps to implementing an effective corrosion control program

Step Discussed in Chapter(s)
• Develop an understanding of the general concepts behind internal corrosion
Ch. 1, Ch. 2
and metal release in drinking water distribution systems
• Determine the extent and magnitude of corrosion Ch. 3
• Determine the possible causes of corrosion Ch. 2, Ch. 3
• Assess corrosion control alternatives Ch. 4
• Develop a corrosion control strategy Ch. 4
• Implement a corrosion control program Ch. 5
• Monitor the effectiveness of the corrosion control program Ch. 3, Ch. 6
• Optimize the corrosion control program if necessary Ch. 5, Ch. 6
M58 book.indb 1 11/17/2010 4:24:30 PM

2 Internal Corrosion Control in Water Distribution Systems
Selection of an effective corrosion control strategy is not a one-time event. Changes in

source water quality, treatment, or distribution operational practices may require a system
to reevaluate, revise, and reimplement a corrosion control program. To assist the reader in
navigating the process described in Table 1-1, this manual is organized accordingly.
Purpose of This Manual__________________________________

Internal corrosion of drinking water distribution systems and home plumbing systems
has long been a recognized issue facing the drinking water industry—one that has many
health, water quality, and economic implications.
This manual was written with the intent to provide a practical overview of internal
corrosion issues, to identify appropriate corrosion response and control methods, and
to develop corrosion control monitoring programs. It is not intended to provide detailed
summaries of corrosion chemistry. References to additional resources are given where the
reader might find more detailed information useful.
Although this manual briefly touches on internal corrosion of nonmetallic pipe sur-
faces, such as asbestos–cement and cement mortar–lined ductile or cast-iron pipe, it pri-
marily focuses on corrosion of metal pipe surfaces, solders, and plumbing fixtures, such
as those composed of lead, copper, and iron. For the purposes of this manual, the term
corrosion refers not only to the electrochemical phenomenon that causes metal loss from
pipe surfaces but also to the dissolution of existing pipe scales and corrosion by-products.
Internal corrosion—that is, corrosion on the interior surface of metal pipes and fixtures—
is the focus of this manual because of the potential for metal release to adversely impact
distributed water quality.
Customer and Infrastructure Impacts of Internal

Corrosion_ ______________________________________________
Internal corrosion can cause degradation of water quality, infrastructure performance
and structural failures, and scaling and reequilibration issues. It may have substantial
economic impacts on water utilities and consumers alike. This section discusses the most
common issues resulting from internal corrosion of drinking water distribution and home
plumbing systems.
Water Quality Deterioration

Internal corrosion of distribution system piping and home plumbing may cause several
water quality problems, including potential health concerns, discoloration, and taste-
and-odor issues. These problems are primarily the result of corrosion of metal pipe sur-
faces, pipe solder, and plumbing fixtures or dissolution of existing pipe scales, although
some problems may be attributable to corrosion of nonmetallic system components (e.g.,
increases in asbestos concentrations).
Health concerns. The majority of the health concerns associated with internal
corrosion are related to the release of trace metal concentrations (e.g., lead, copper, cad-
mium, and so on) from corroding metal surfaces. The potential health concerns related to
increased metal concentrations in drinking water are discussed further in the regulatory
section in this chapter.
Color. Many of the color issues typically encountered in drinking water are attrib-
utable to internal corrosion (Kirmeyer et al. 2000). Corrosion of cast-iron pipe or dissolu-
tion of existing scale on cast-iron pipe may result in rust-colored water or red water due
to the presence of ferric iron (Fe(III)). Red water can stain laundry and plumbing fixtures.
Iron corrosion may also result in yellow or black water, which is a result of the presence
M58 book.indb 2 11/17/2010 4:24:31 PM

Overview of Internal Corrosion Impacts in Drinking Water Distribution Systems 3
Courtesy of District of Columbia Water (DC Water).
Figure 1-1 Red water sample from a US distribution system
of ferrous iron (Fe(II)). Ferrous iron may occur in waters with low dissolved oxygen con-
centrations. Black water may also be the result of the presence of manganese in source
waters. Copper corrosion may result in blue water, which can also stain bathroom fixtures
and hair. Gray or black water may also be attributable to hot water heaters. Corrosion of
zinc in galvanized piping may result in a milky appearance.
Red water (Figure 1-1) may occur as a result of “hydraulic entrainment”—that is,
the suspension and transport of loose corrosion deposits resulting from changes in flow
or direction of flow—or “iron uptake”—that is, the release of ferrous ions at the pipe sur-
face and subsequent oxidation to ferric iron and precipitation in the bulk water (Smith et
al. 1998). Alternating periods of stagnation, anaerobic conditions, and temperature have
been observed to cause red water, as have significant changes in alkalinity and chlorine
residual (oxidation–reduction potential [ORP]). Maintaining consistent water quality and
maintaining conditions that produce a hard iron scale are critical to minimizing the poten-
tial for red water (Smith et al. 1998; Reiber 2006).
Blue water may be the result of the presence of either dissolved or particulate copper
corrosion by-products. Dissolved copper by-products are primarily the result of low pH
and may usually be eliminated by raising pH to above 7.5 (Edwards et al. 2000). However, it
is often more difficult to determine the cause of particulate copper corrosion by-products,
which are found in many home plumbing systems. Particulate copper corrosion by-prod-
ucts are thought to be the result primarily of microbiological, water quality, or physical
factors or some combination of factors that frequently result in pitting corrosion and/or
pipe failure in addition to blue water (Edwards et al. 2000; Bremer et al. 2001).
Taste and odor. Dissolved and particulate iron may cause taste and odor at low con-
centrations. The presence of iron has been shown to result in metallic taste (Khiari et al.
2002). In addition to metallic taste, cast-iron corrosion may also result in musty tastes and
odors (Kirmeyer et al. 2000). Aluminum and zinc may contribute to an astringent mouth-
feel, and zinc may also result in a sour taste. Reactions of disinfectants with cement–
mortar linings in ductile-iron pipe may also result in astringent, oily (rancid), pine, and
phenolic odors (Khiari et al. 2002). High pH may cause the release of phenols from corrod-
ing asbestos–cement pipe, which can react with chlorine to form chlorophenols (Kirmeyer
et al. 2000).
M58 book.indb 3 11/17/2010 4:24:32 PM

Infrastructure Impacts
In addition to affecting water quality, internal corrosion may also impact distribution and
plumbing infrastructure. Most notably, internal corrosion may result in failures of home
plumbing systems, causing extensive property damage. Deterioration of distribution sys-
tem piping and valves as well as deterioration of meters and other in-line devices may also
result from corrosion.
Home plumbing failures. The most significant home plumbing failures that occur
as a result of internal corrosion are copper pitting and pinhole leaks. Pitting corrosion,
though not well understood, not only damages copper plumbing but the associated leak-
age may also cause substantial damage to homes and result in mold growth and other
issues that may pose health concerns (Edwards 2004). Table 1-2 summarizes the condi-
tions under which potential copper corrosion has been traditionally thought to occur.
In addition to pitting corrosion, microbially influenced corrosion (MIC) of copper
piping may also cause pitting and failure of home plumbing (Cantor et al. 2003; Cantor et
al. 2006). In fact, it is believed that MIC is an often overlooked contributor to corrosion of
plumbing systems. It has been demonstrated that MIC frequently may occur in plumbing
systems with long stagnation times, in sulfide-containing waters, and in areas in which
there is little to no disinfectant residual remaining (Bremer et al. 2001; Jacobs et al. 1998;
Cantor et al. 2003).
Corrosion of distribution system piping may lead to pipe and valve failures resulting
in increased water losses. In areas of low or negative pressure, inflow and infiltration are
possible, as are bacterial contamination and other contamination from surrounding soils
and groundwaters. Failure of corroded valves may also cause operational issues, such
as failure to isolate water mains in the event of a line break, loss of pressure, negatively
impact the ability to conduct distribution flushing, and creation of stagnant areas or areas
of low flow in the distribution system (e.g., if a valve stem breaks in the closed position).
Failures of meters and other in-line devices may reduce system revenues because of inac-
curate meter readings and may minimize the ability to accurately determine water loss or
identify main breaks.
Scaling and Reequilibration

Scaling and reequilibration are two key factors that may impact corrosion and system
performance. Formation of protective metal scales is the primary corrosion control mech-
anism regardless of the technology employed. For example, when pH and alkalinity adjust-
ment are used as the primary lead control technology, the objective is to form metal solids
(e.g., lead carbonate and lead oxides) on the pipe surface and thereby prevent corrosion,
or dissolution, of lead. Similarly, the objective of orthophosphate addition is the formation
of a lead phosphate layer (Vik et al. 1996).
Controlling the type of scale formed and subsequently maintaining the stability of
those scales are key to an effective corrosion control program. Iron scales, for example,
are present in either the ferrous (Fe2+) form or ferric (Fe3+) form. Ferrous iron scales are
much softer than ferric iron scales and are more likely to contribute to red water (Reiber
2006). Creating conditions in which harder ferric iron scales are formed and maintained is
ideal to maintaining water quality.
After a protective scale is formed in distribution piping or home plumbing, maintain-
ing scale stability is essential to effective corrosion control treatment. Changes in distri-
bution system water quality may result in reequilibration of existing scales causing red
water, increased tap lead or copper concentrations, or other issues.
For example, in 2002, the Washington Aqueduct switched from free chlorine to
chloramine to reduce disinfection by-product (DBP) concentrations in its customer agen-
cies’ systems. As a result of this change, the ORP of the finished water changed, causing
M58 book.indb 4 11/17/2010 4:24:32 PM

Table 1-2 Summary of potential copper corrosion issues

Type of Corrosion
Characteristic Type I Pitting Type II Pitting Type III Pitting
of Corrosion Uniform Corrosion (Cold Water) (Hot Water) (Soft Water)
Narrower than
Pit shape No pits Deep and narrow Type I Wide and shallow
Type of problem Blue or green water, Pipe failure Pipe failure Blue water, volumi-
present with cor- high by-product re- nous by-product re-
rosion lease leases, pipe blockage
Scale morphol- Tarnished copper Underlying Cu2O with Underlying Cu2O Underlying Cu2O
ogy on attacked surface or loose overlying malachite, with overlying bron- with overlying bron-
surface powdery scale calcite, or other basic chantite, some mala- chantite, some mala-
copper salts, occasion- chite chite
ally CuCl underlies
Cu2O
Water quality Soft waters of low Hard, cold well waters Hot waters, pH be- Soft waters, pH >8.0
pH (<7.2) between pH 7.0 and 7.8, low 7.2, high sulfate
high sulfate relative to relative to bicarbon-
chlorides and bicarbonate, occasional Mn
ate, high CO2 deposits
Initiating factors None noted Stagnation early in pipe Higher tempera- Stagnation early in
life, deposits within tures, high chlorine pipe life, pH >8.0,
pipe include dirt or car- residuals, alum co- alum coagulation,
bon films, high chlorine agulation, particles low chlorine residual
residuals, water soften-
ers, alum coagulation
Ameliorating Raise pH or increase NOM, increase bicar- Lower temperatures, NOM, avoid stagna-
factors and treat- bicarbonate bonate and pH higher pH, increase tion early in pipe
ments bicarbonate and pH life, increase hard-
ness and alkalinity,
elevate chlorine re-
sidual to >0.5 mg/L
NOM = natural organic matter
Source: Edwards et al. 1994.
a shift in the speciation of the existing lead scales and elevated lead concentrations in the
District of Columbia Water (DC Water) service area (USEPA 2007). It should be noted that
the conversion to chloramine was just one factor that contributed to the lead release. Many
operating chloraminated systems have not experienced similar results. Refer to Appendix
B for more details of the Washington, D.C., project.
Economic Issues
Internal corrosion may have a number of economic or cost impacts. Failure of distribution
system piping and home plumbing may necessitate costly repairs, not only of failed piping
but also of other assets damaged as a result of the pipe failure. There are also more subtle
economic issues associated with corrosion. Increased lead and copper levels may not only
result in negative publicity but may also have the potential to result in litigation. The costs
associated with legal action may be substantial. Water quality issues resulting from cor-
rosion (e.g., color, taste and odor, and lead) may also cause negative customer perceptions
and not only impact (reduce) usage but also cost utilities potential future customers.
M58 book.indb 5 11/17/2010 4:24:32 PM

Regulatory Impacts______________________________________
Lead and Copper
The Lead and Copper Rule (LCR) was promulgated in 1991 with the purpose of reducing
drinking water exposures to lead and copper (56 FR 26460) (Federal Register 1991). The
purpose of the LCR is to protect public health by minimizing lead and copper levels in drink-
ing water, primarily by reducing water corrosivity (USEPA 2004a). Lead has been demon-
strated to cause delays in physical and mental development in infants and children and has
been linked to deficits in attention span and learning abilities (USEPA 2006a). Short-term
copper exposure may cause gastrointestinal distress; long-term exposure may cause liver
or kidney damage. In individuals with Wilson’s disease, which causes the body to retain
copper, copper can cause severe brain damage, liver failure, and death (NIH 2006).
USEPA estimates that approximately 20 percent of human lead exposure is from lead
in drinking water (USEPA 2006a). The LCR established a maximum contaminant level
goal (MCLG) of zero for lead and an MCLG of 1.3 mg/L for copper. The rule established a
National Primary Drinking Water Regulation for lead and copper that consisted of a treat-
ment technique requirement including corrosion control treatment, source water treat-
ment, lead service line (LSL) replacement, and public education. The rule set an action
level (AL) of 0.015 mg/L for lead and an AL of 1.3 mg/L for copper. If the 90th percentile
concentration for lead or copper is above the AL, public water systems may be required to
initiate water quality parameter monitoring for key corrosion control parameters, install
corrosion control treatment, begin source water monitoring or treatment, replace LSL, or
undertake a public education program.
Internal corrosion of lead service lines, brass meters and plumbing fixtures, and cop-
per plumbing contributes to the concentrations of lead and copper in drinking water. Pure
lead pipe (lead service lines), lead solder, and brass with greater than 8 percent lead were
banned by the 1986 Safe Drinking Water Act Amendments (Dudi et al. 2005). “Lead-free”
brass, however, may contain as much as 8 percent lead by weight. Greenville Utilities in
Greenville, N.C., exceeded the lead AL in 2004 and 2005 despite the fact that the system
does not contain any lead service lines. In this case, the exceedance was attributed to
leaching from lead-based solder and plumbing fixtures containing lead (Landers 2006).
California banned the use of pipe solder with lead content greater than 0.2 percent in 1987.
On Jan. 1, 2010, a new California law went into effect defining “lead free” as less than 0.25
percent. All pipes, fittings, and fixtures must be lead free.
Impact of the LCR on tap-water lead concentrations. A review conducted by
USEPA of LCR compliance monitoring data for public water systems serving more than
3,300 people shows that since 2000, fewer than 4 percent of those systems have exceeded
the lead AL (USEPA 2004b). Prior to 2002, states were not required to report 90th percen-
tile lead concentrations to USEPA unless those values exceeded the AL. Consequently,
it is difficult to compare the tap-water lead concentrations immediately following imple-
mentation of the LCR to current levels. However, USEPA did evaluate data from 166
large public water systems that exceeded the AL following initial LCR monitoring in
1992 and 1993 (USEPA 2006f). Following monitoring conducted by these same utilities
between 2000 and 2004, only 15 of those systems continued to exceed the AL (Figure 1-2).
A closer look at the data presented in Figure 1-2 reveals tap-water lead concentra-
tions have decreased significantly in those systems that initially exceeded the lead AL.
Figure 1-3 shows that, for those 166 large systems that exceeded the AL in 1992–1993,
the average 90th percentile lead concentration decreased from nearly 32 µg/L to 8.2 µg/L.
Over that same period, the maximum 90th percentile lead concentration decreased from
211 µg/L to 84 µg/L.
M58 book.indb 6 11/17/2010 4:24:32 PM

Below Action Level Above Action Level

200
175 166
151
150
Number of Systems
125
100
75
50
25 15
0
0
1992/93 Most Recent Data (2000-2004)
Time Period for Monitoring

Source: USEPA 2006f.
Figure 1-2 Number of large US water systems exceeding the lead action level
120%
1992/93
2000-2004
100%
80%
Percentile
60% 1992/93 2000-2004

Average 31.9 8.2
Minimum 2 0
40% 10th Percentile 11 1
25th Percentile 17.3 3
Median 23 6
20% 75th Percentile 36 11
90th Percentile 62 15
Maximum 211 84
0%
0 50 100 150 200 250
90th Percentile Lead Concentration (µg/L)
Figure 1-3 Comparison of 90th percentile LCR monitoring data for 166 large public water systems
Revisions to the Lead and Copper Rule. USEPA revised the LCR in 2007 (72 FR
57781, [Federal Register 2007]) with the intent of enhancing implementation in the areas
of monitoring, treatment, customer awareness, and lead service line replacement. In addi-
tion, the 2007 revisions intended to improve public education by ensuring drinking water
customers receive “meaningful, timely, and useful information needed to help them limit
their exposure to lead in drinking water.”
M58 book.indb 7 11/17/2010 4:24:33 PM

The most significant revisions to the LCR clarify some of the confusion regarding
sample collection and reporting. Specifically, the LCR revisions require that all compliance
monitoring results, including those above the required number of samples, be included in
the 90th percentile determination and that all samples be taken within the same calendar
year. The remainder of the revisions address customer notification for those residents in
a utility’s sampling program and provisions to deal with LSL that were previously “tested
out” due to levels below the AL.
Iron
USEPA established a secondary maximum contaminant level (SMCL) of 0.3 mg/L for iron,
which represents a “reasonable goal for drinking water quality” (40 Code of Federal Regu-
lations [CFR] 143.3). SMCLs control contaminants that primarily affect the aesthetic qual-
ity of drinking water and are not federally enforceable, although state primacy agencies
have the authority to include them in state drinking water regulations. When present at
sufficiently high levels, iron may result in rusty color, deposit of sediment, metallic taste,
and reddish or orange staining.
Much of the iron present in water is the result of natural mineral deposits and, when
present in source waters at concentrations above the SMCL, iron is frequently removed at
the water treatment facility. If not removed, source-water iron may result in red water in
the distribution system. However, internal corrosion of unlined cast-iron pipe, galvanized
pipe, or dissolution of existing pipe scales in iron pipe may also result in aesthetic issues
in drinking water distribution systems.
Cadmium
Short-term exposure to high concentrations of cadmium may result in nausea, vomit-
ing, diarrhea, muscle cramps, salivation, sensory disturbances, liver injury, convulsions,
shock, and renal failure. Long-term exposures may cause kidney, liver, bone, and blood
damage (USEPA 2006b). Consequently, USEPA has established an MCLG of 0.005 mg/L for
cadmium as well as a maximum contaminant level (MCL) of 0.005 mg/L.
Cadmium is primarily present in drinking waters due to erosion of natural depos-
its but may also be present due to leaching from cement–mortar lined pipes (Guo et al.
1998; Berend and Trouwborst 1999) or as a result of corrosion of galvanized piping. It has
also been observed that the amount of lead and cadmium released from galvanized piping
decreased with the age of the pipe due to the formation of a passivating layer on the pipe
surface (Meyer 1980).
Zinc
Zinc in drinking water is primarily an aesthetic concern as it may result in a metallic
taste. As a result, USEPA has established an SMCL of 5 mg/L for zinc. Zinc may be pres-
ent in drinking water due to erosion of natural deposits or as a result of dezincification of
brass plumbing fixtures. Several studies of low-alkalinity waters have found that iron and
zinc were the principal corrosion by-products of galvanized plumbing (Dangel 1975; J.M.
Montgomery 1982).
The zinc coating on galvanized pipe may contain lead, copper, cadmium, chromium,
aluminum, barium, and other impurities. There are a number of standards regarding gal-
vanized pipe, most notably by the American Society for Testing and Materials (ASTM) in
the United States. As a result of these impurities, corrosion of galvanized pipe may result
in the release of trace metal concentrations.
M58 book.indb 8 11/17/2010 4:24:33 PM

Asbestos
Asbestos in drinking water is most commonly present as a result of degradation of
asbestos–cement piping used in drinking-water distribution systems. Though it may be
present in natural soils, it does not often migrate to groundwater through soils (USEPA
2006c). Short-term exposure to asbestos in drinking water is not known to cause any
health issues; long-term exposure may result in increased risk of developing benign
intestinal polyps. USEPA has established an MCLG of 7 million fibers per liter (M.L.) and
an MCL of 7 M.L. for asbestos.
Biological Regrowth and Chlorine Demand

The Surface Water Treatment Rule was finalized in 1989 and requires maintenance of a
disinfectant residual in the distribution system, among other requirements (54 FR 27486
[Federal Register 1989a]). The Total Coliform Rule (TCR) imposes monitoring require-
ments for total coliform, as well as for fecal coliform and Escherichia coli (54 FR 27544
[Federal Register 1989b]). A Revised Total Coliform Rule “Agreement in Principle” was
signed in September 2008, which will change the requirements associated with the rule.
Biological regrowth and chlorine demand may be impacted by corrosion in a number of
ways and vice versa. Chlorine and other oxidants impact the ORP, which is a factor in
maintaining effective corrosion control. Conversely, corrosion by-products can exert an
oxidant demand, diminishing distribution system disinfectant residuals and resulting in
increased potential for microbiological regrowth. Microbiological regrowth not only may
threaten compliance with the TCR but also may result in increased potential for MIC to
occur.
Release of Trace Metals from Cementitious Coatings

Barium, cadmium, chromium, and aluminum have been found to leach from cement–
mortar lining in distribution piping (Guo et al. 1998; Berend and Trouwborst 1999). While
this leaching represents a potential water quality and public health concern, it is not the
focus of this manual.
ADDITIONAL READING______________________________________
The reader is advised to become more knowledgeable in the following areas:
• The water quality and operational factors that influence internal corrosion and
metal release in drinking water distribution systems
• The potential water quality, health-related, infrastructure, and regulatory con-
cerns associated with internal corrosion
• The secondary impacts associated with changes in source-water quality, treatment,
or distribution operations and the potential for those changes to impact existing
corrosion control effectiveness (i.e., reequilibration of existing pipe scales)
References for this chapter give detailed information on these topics and the other
topics discussed. The reader is encouraged to follow up on the referenced material that is
publicly available to obtain keener insight into this discussion.
In addition to the references, the following resources provide substantial discussions
regarding these topics and will be extremely valuable to the reader in developing an under-
standing of the factors that influence corrosion, implementing an effective corrosion treat-
ment and monitoring program, and assessing the cause of future corrosion-related water
quality issues.
M58 book.indb 9 11/17/2010 4:24:33 PM

AwwaRF and TZW, Internal Corrosion of Water Distribution Systems, 2nd ed. Denver,
Colo.: AwwaRF and AWWA 1996.
G.J. Kirmeyer, G. Pierson, J. Clement, A. Sandvig, V. Snoeyink, W. Kriven, and A. Camper.
Distribution System Water Quality Changes Following Corrosion Control Strate-
gies. Denver, Colo.: AwwaRF and AWWA 2000.
G. Kirmeyer, B. Murphy, and A. Sandvig. Post Optimization Lead and Copper Control
Monitoring Strategies. Denver, Colo.: AwwaRF and AWWA 2004.
T. Case. Distribution System Corrosion and the Lead and Copper Rule: An Overview of
AwwaRF Research. Denver, Colo.: AwwaRF and AWWA 2004.
AWWA. Managing Change and Unintended Consequences: Lead and Copper Rule Cor-
rosion Control Treatment. Denver, Colo.: AWWA 2005.
REFERENCES________________________________________________
Berend, K., and T. Trouwborst. 1999. Cement– Federal Register. 1989a. National Primary
Mortar Pipes as a Source of Aluminum. Drinking Water Regulations; Filtra-
Jour. AWWA, 91(7):91–100. tion, Disinfection; Turbidity, Giar-
Bremer, P.J., B.J. Webster, and D.B. Wells. 2001. dia lamblia, Viruses, Legionella, and
Biocorrosion of Copper in Potable Water. Heterotrophic Bacteria, Final Rule.
Jour. AWWA, 93(8):82–91. 54(124):27486–27567.
Cantor, A.F., J.B. Bushman, and M.S. Glodoski. Federal Register. 1989b. National Primary
2003. A New Awareness of Copper Pipe Drinking Water Regulations; Total Coli-
Failures in Water Distribution Systems. In forms (Including Fecal Coliform and E.
Proc. of the AWWA Water Quality Tech- coli); Final Rule. 54(124):27544–27568.
nology Conference. Denver, Colo.: AWWA. Federal Register. 1991. National Primary
Cantor, A.F., J.B. Bushman, M.S. Glodoski, E. Drinking Water Regulations; Lead and
Kiefer, R. Bersch, and H. Wallenkamp. Copper; Final Rule. 56(110):26460.
2006. Copper Pipe Failure by Microbio- Federal Register. 2007. National Primary
logically Influenced Corrosion. Materials Drinking Water Regulations for Lead
Performance, 45(6):38. and Copper: Short-Term Regula-
Dangel, R.A. 1975. Study of Corrosion Products tory Revisions and Clarifications,
in the Seattle Water Department Distribu- 72(195):57781–57820.
tion System. Report No. 670/2-75-036. Cin- Guo, Q., P.J. Toomuluri, and J.O. Eckert Jr. 1998.
cinnati, Ohio: USEPA. Leachability of Regulated Metals From
Dudi, A., M. Schock, N. Murray, and M. Edwards. Cement–Mortar Linings. Jour. AWWA,
2005. Lead Leaching From Inline Brass 90(3):62–73.
Devices: A Critical Evaluation of the Exist- Jacobs, S., S. Reiber, and M. Edwards. 1998.
ing Standard. Jour. AWWA, 97(8):66–78. Sulfide-Induced Copper Corrosion. Jour.
Edwards, M. 2004. Written testimony to Chair- AWWA, 90(7):62–73.
man Tom Davis and U.S. House of Rep- J.M. Montgomery Consulting Engineering Inc.
resentatives, Committee on Government (JMM). 1982. Internal Corrosion Mitiga-
Reform, May 5, 2004. tion Study, Final Report. Report prepared
Edwards M. 2006. A Perspective on Premise for the Portland, Ore., Water Bureau.
Plumbing and Water Quality Degradation. Khiari, D., S. Barrett, R. Chinn, A. Bruchet, P.
In Proc. of the AWWA Annual Conference Piriou, L. Matia, F. Ventura, I. Suffet, T.
and Exposition. Denver, Colo.: AWWA. Gittelman, and P. Leutweiler. 2002. Dis-
Edwards, M., J.F. Ferguson, and S.H. Reiber. tribution Generated Taste and Odor
1994. The Pitting Corrosion of Copper. Phenomena. Denver, Colo.: AwwaRF and
Jour. AWWA, 86(7):74–90. AWWA.
Edwards, M., S. Jacobs, and R.J. Taylor. 2000. Kirmeyer, G.J., M. Friedman, and J. Clement.
The Blue Water Phenomenon. Jour. 2000. Guidance Manual for Maintain-
AWWA, 92(7):72–82. ing Distribution System Water Quality.
Denver, Colo.: AwwaRF and AWWA.
Landers, J. 2006. Getting the Lead Out. Public
Works, 137(4):30–32.
M58 book.indb 10 11/17/2010 4:24:33 PM

Meyer, E. 1980. Beeinträchtigung der USEPA. 2005. Medium and Large Public Water
Trinkwassergüte durch Anlagenteile Systems Exceeding the Action Level Sum-
der Hausinstallation—Bestimmung des mary from SDWIS/FED as of January 27,
Schwermetalleintrags in das Trinkwasser 2005. http://www.epa.gov/safewater/lcrmr/
durch Korrosionsvorgänge in metallischen pdfs/summary_lcmr_sdwisfed_data.pdf.
Rohren. DVGW-Schriftenreihe Wasser USEPA. 2006a. Lead in Drinking Water: Basic
Nr., 23:113–131. Information. http://www.epa.gov/safewater/
National Institutes of Health (NIH). 2006. Wil- lead/basicinformation.html.
son Disease. National Digestive Diseases USEPA. 2006b. Consumer Fact Sheet on: Cad-
Information Clearinghouse, National Insti- mium. http://www.epa.gov/OGW DW/con-
tute of Diabetes and Digestive and Kidney taminants/dw_contamfs/cadmium.html.
Diseases, National Institutes of Health. USEPA. 2006c. Consumer Fact Sheet on: Asbes-
http://digestive.niddk.nih.gov/ddiseases/ tos. http://www.epa.gov/OGW DW/contam-
pubs/wilson/index.htm. inants/dw_contamfs/asbestos.html.
Reiber, S. 2006. Corrosion, Red Water, and Dis- USEPA. 2006d. Consumer Fact Sheet on: Cop-
tribution System Water Quality. In Proc. of per. http://www.epa.gov/OGWDW/contam-
the AWWA Annual Conference and Expo- inants/dw_contamfs/copper.html.
sition. Denver, Colo.: AWWA. USEPA. 2006e. Drinking Water Standards.
Smith, S.E., T. Ta, D.M. Holt, A. Delanoue, and J. http://www.epa.gov/OGWDW/standards.
Colbourne. 1998. Minimizing Red Water in html.
Drinking Water Distribution Systems. In USEPA. 2006f. Lead 90th Percentile Levels for
Proc. of the AWWA Water Quality Tech- 166 Large Water Utilities—Then and Now.
nology Conference. Denver, Colo.: AWWA. ht t p://w w w.epa .gov/sa fewater/ lcr m r/
US Environmental Protection Agency (USEPA). reductionplan_comparrison.html.
1991. Safe Drinking Water Act as Amended USEPA. 2006g. Lead and Copper Rule State
by the Safe Drinking Water Act Amend- File Review: National Report. EPA 816-R-
ments of 1986. Report No. 570/B-91-042. 06-001. Washington, D.C.: USEPA, Office
Washington, D.C.: USEPA Office of Water. of Water.
USEPA. 2002. Health Risks from Microbial USEPA. 2007. Elevated Lead in DC Drinking
Growth and Biofilms in Drinking Water Water. EPA 815-R-07-021. Washington,
Distribution System. Washington, D.C.: D.C.: USEPA, Office of Water.
USEPA Office of Ground Water and Drink- Vik, E.A., R.A. Ryder, I. Wagner, and J.F. Fergu-
ing Water. son. 1996. Mitigation of Corrosion Effects.
USEPA. 2004a. Lead and Copper Rule: Quick In Internal Corrosion of Water Distri-
Reference Guide. EPA 816-F-04-009. Wash- bution Systems. 2nd ed. Denver, Colo.:
ington, D.C.: USEPA, Office of Water. AwwaRF and AWWA.
USEPA. 2004b. Summary of Lead Action Level
Exceedances for Medium (3,300–50,000)
and Large (>50,000) Public Water Systems.
http://www.epa.gov/safewater/lcrmr/pdfs/
data_lcmr_leadsummary_0522041.pdf.
M58 book.indb 11 11/17/2010 4:24:34 PM

M58 book.indb 12 11/17/2010 4:24:34 PM

AWWA MANUAL M58
Chapter 2
Fundamentals of Internal
Corrosion and Metal
Release
Abigail F. Cantor, P.E.

Process Research Solutions LLC
Introduction_____________________________________________
There are many misunderstandings about corrosion in water distribution systems. Just
using the word corrosion sets us on the wrong track. We are actually concerned about
elevated metal concentrations in drinking water. Corrosion, especially uniform corrosion
that is typically the focus of corrosion control studies, is only one of several mechanisms
that can elevate the metal concentrations (Table 2-1).
This chapter will describe the common mechanisms that may cause metal release
into water and, thereby, elevate metal concentrations. Understanding and managing these
mechanisms are the keys to achieving what is commonly called corrosion control.
Mechanisms of Metal Release by Uniform Corrosion____

The Need for a Protective Barrier
The correct way to apply the term corrosion to drinking water systems is in reference to
an electrochemical interaction between a metal surface, such as a pipe wall, and the water
with which it is in contact. The electrochemical interaction is similar to that of a battery
(Figure 2-1).
The components of a battery are
• Negative terminal called the anode
• Positive terminal called the cathode
13
M58 book.indb 13 11/17/2010 4:24:34 PM

Table 2-1 Chapter 2 key points

• Uniform corrosion, which is typically the focus of corrosion control studies, is only one of many mechanisms
that can cause the release of lead, copper, and other metals into water.
Uniform Corrosion
• In uniform corrosion, the anodic and cathodic sites move around dynamically on the pipe wall, resulting in
uniform loss of metal.
• Calcium carbonate has been shown to be inadequate as a protective barrier against uniform corrosion.
• Naturally occurring constituents such as dissolved inorganic carbon (DIC), hydroxides, and oxygen can form
compounds with dissolved metals. Those compounds create protective barriers against uniform corrosion
to varying degrees. The adequacy of the pipe films as protective barriers depends on the solubility of the
compounds formed.
• Chemicals, such as orthophosphate, can be added to produce adequate barriers against uniform corrosion on
the pipe wall.
Nonuniform Corrosion
• In nonuniform corrosion, the anode remains in one place. Metal is lost from that localized point.
• There are many types of nonuniform corrosion.
• For each type of nonuniform corrosion, pitting and mounded corrosion debris may or may not be observed.
• Microbially influenced corrosion, one type of nonuniform corrosion, needs to be considered routinely as it may
play a bigger role in metal release than realized.
Other Metal Release
• There are other mechanisms of metal release outside of corrosive processes that should not be overlooked.
Adsorption and release of metal by chemical scales are important examples.
• Degrading tank liners and equipment gaskets may significantly add to increased metal in the water.
Summary
• Understanding and addressing the various mechanisms of metal release into water in addition to uniform
corrosion are the keys to achieving what is commonly called corrosion control.
• Medium to carry electrons from the anode to the cathode

• Chemical that can accept the electrons at the cathode
• Chemical that can pair with the metal ion released at the anode
In uniform corrosion taking place inside the piping of water distribution systems, the
anode and cathode occur dynamically at random sites on the pipe wall (Figure 2-2). These
“negative and positive terminals” are microscopic in surface area. They are thought to
occur because of natural microscopic imperfections and variations on the metal surface
(AwwaRF and TZW 1996).
The electrons flow from the microscopic anodes to the microscopic cathodes through
the pipe wall, which is similar to the role that wiring plays in an electrical circuit.
The water solution in contact with the cathode provides the chemicals to accept the
electrons. For instance, oxygen in the water solution can accept an electron and the reduc-
tion reaction occurs:
e – + ¼ O2 + ½ H2O  OH–
In the meantime, metal atoms from the solid metal at the anode, having given up elec-
trons, undergo an oxidation reaction. The oxidized metal is now soluble in the adjacent
water solution. That is, solid metal has been lost, transferred to water as a metal ion. It is
this oxidation of the metal that is termed “corrosion.”
M58 book.indb 14 11/17/2010 4:24:34 PM

Fundamentals of Internal Corrosion and Metal Release 15
e–
Cathode Anode
C Zn
Zn++
2e–
ZnSO4
H2 (gas)
H2SO4 2H+ and SO4––
Figure 2-1 Typical battery
Reduction rxn at the cathode: e– + ¼ O2 + ½ H2O = OH–

Oxidation rxn at the anode: Me(0) = 2e– + Me++
H2 0
O2
OH –
Me(OH)2
2e– O2 Me++ O2
+ – –
2e– + + + 2e–
–
Metal Pipe
Figure 2-2 Typical water and metal pipe “battery”
The water solution now contains positively charged metal ions from the anode and
negatively charged ions from the cathode. New compounds are formed as oppositely
charged ions pair.
Figure 2-2 illustrates methods to stop uniform corrosion from occurring in a simpli-
fied manner. That is, because there is no wire with an on/off switch to break the flow of
electrons, only two components can be manipulated. First, electron acceptors in the water
solution, such as oxygen, can be eliminated. This approach is sometimes but not always
practical in water treatment. A second approach is to create a barrier, or protective film,
between the metal surface and the water solution (Figure 2-3). However, the film must be
an insoluble, uniform, nonporous layer that prevents the electron acceptors from contact-
ing the electrons at the cathodic sites on the pipe wall.
The Failure of Calcium Carbonate Scales as Protection Against

Uniform Corrosion
It was stated that if a barrier can be created between the metal surface and the water,
uniform corrosion can be stopped. But, it was also stipulated that the barrier must at least
be uniform and nonporous to prevent contact between the metal and the water. Many
decades ago, it was thought that calcium carbonate could create such a barrier.
M58 book.indb 15 11/17/2010 4:24:35 PM

H2 0
O2
O2 O2
+ – –
+ + +
–
Metal Pipe
Figure 2-3 Barrier between water and metal pipe
Calcium carbonate naturally occurs in water and is easy to add to water and manipu-
late in water. It is known that adjustments to pH, temperature, calcium hardness, and
alkalinity can force calcium carbonate to come out of solution. A number of simple meth-
ods have been developed to calculate relative quantities of calcium carbonate that will
precipitate as a scale. The most common of these methods are the Langelier Index, the
Calcium Carbonate Precipitation Potential, the Aggressiveness Index, and the Ryznar
Index. The goal has been to achieve a certain quantity of calcium carbonate scale that will
stop uniform corrosion but not so much scale that smaller pipes will become significantly
reduced in diameter or plugged.
It would be convenient if corrosion control and control of metal release were as sim-
ple as calculating the Langelier Index. However, an important corrosion reference book,
Internal Corrosion of Water Distribution Systems (AwwaRF and TZW 1996), emphasizes
the inadequacies of calcium carbonate scale. These arguments are based on microscopic
studies of this scale’s physical properties on lead, copper, and iron pipe walls (AwwaRF
and TZW 1996). It has been confirmed that this coarse scale cannot stop the electron recep-
tor molecules from migrating to the pipe wall. Studies cited in the book also show the lack
of correlation of Langelier Index with concentration of metals in the water (AwwaRF and
TZW 1996).
If calcium carbonate is inadequate, what can create a physical barrier on a metal
surface to prevent the transmission of electron acceptors to the cathodic sites?
Other Pipe Films and Scales

The following section will review how a number of other substances are related to uniform
corrosion including lead, copper, copper alloys, iron, galvanized iron pipe, and cement-
lined pipe.
Lead
Dissolved inorganic carbon and other naturally occurring ions. Researchers began
their exploration for adequate protective barriers on lead pipe by examining the structure
and composition of scales that had built up naturally on pipe walls. A classic example of
this type of exploration was described in a 1987 article by Colling and colleagues. The
researchers examined the composition of lead pipe wall deposits using X-ray diffraction
(XRD) and infrared spectroscopy. The deposits’ structure was viewed with scanning elec-
tron microscopy (SEM). Other similar studies were performed in Europe and in the United
States in the 1970s and 1980s (AwwaRF and TZW 1996).
The most common compounds found on a lead pipe wall are cerussite (PbCO3) and
hydrocerussite (Pb3(CO3)2)(OH)2). Researchers observed that “the concentrations of lead
M58 book.indb 16 11/17/2010 4:24:35 PM

[in the water] are controlled by the solubility of lead salts in the pipe deposits” (Colling et
al. 1987).
Another way to look at deposition on lead pipe is to think about the metal ions
released at the anodic sites on the pipe wall (Figure 2-2). Those metal ions form com-
pounds with other ions already in the water. These compounds are referred to as corro-
sion by-products. Each corrosion by-product has its own solubility in water. When any
compound exceeds the saturation concentration for its solubility, excess compound pre-
cipitates onto the pipe wall as a solid. If that solid has a low solubility and also forms a
uniform and nonporous layer, an adequate barrier is created to diminish and even prevent
further metal release into the water. This barrier breaks the circuit of uniform corrosion.
Given that lead pipe deposits were found to have carbonate (CO3 –2) compounds as a
major component, dissolved inorganic carbon (DIC) became the focus of subsequent stud-
ies. Controlled experiments with lead pipes were devised. DIC and pH were varied in the
water, and the resultant lead concentration was studied (Colling et al. 1987; AwwaRF and
TZW 1996).
Taking this relationship a step further, a theoretical solubility model was developed.
Thermodynamic solubility constants were used to calculate the lead concentration in
water based on DIC and pH (Schock 1980). Various compounds of lead combined with
carbonate, hydroxide, and oxygen were taken into consideration in this model.
In summary, DIC and pH are major factors in determining the concentration of lead
in the water. The concentration of lead in the water is based on the solubility of compounds
that the lead forms with those constituents and other naturally occurring constituents in
water. The structural characteristics of a scale or film as it covers the pipe wall are also
important.
Orthophosphate. So far, this discussion has described protective barriers formed
by naturally occurring ions such as CO3 –2, O –2, and OH–1. There are chemicals that can be
added to the water that may also form protective barriers to minimize uniform corrosion
of lead.
One common chemical is orthophosphate. Its ability to form protective (also called
passivating) films on lead surfaces in contact with water uses the very same mechanisms
previously discussed. That is, metal ions are released into the water from the anodic sites
on the metal surface. The metal ions form compounds (corrosion by-products) with the
phosphate ions that have been added to the water. When the saturation concentration of
any lead/phosphate compound is reached, the compound precipitates out of the water as a
solid and forms a film on the pipe wall and other exposed surfaces. If that compound has
low solubility and has the proper structural characteristics to prevent contact of the pipe
with the water, uniform corrosion is minimized or essentially prevented.
Several lead and phosphate compounds are possible, but research has concluded
that a compound called hydroxypyromorphite (Pb5(PO4)3OH) is most likely the solid that
controls the solubility of lead in the water when orthophosphate has been added (AwwaRF
and TZW 1996). It has also been concluded that the solubility of the lead and orthophos-
phate compounds that form is dependent on pH and DIC in addition to the concentration
of orthophosphate (AwwaRF and TZW 1996).
Polyphosphates and blends. Polyphosphates and polyphosphate–orthophosphate
blends are also used to create corrosion control barriers on pipe walls. Polyphosphates
are polymers composed of repeated linked units of orthophosphate ions in various molec-
ular structures. They have completely different properties than simple nonpolymeric
orthophosphate compounds such as phosphoric acid (H3PO4) or potassium orthophos-
phate (K3PO4). How the phosphate ions in orthophosphate compounds can form new com-
pounds with lead ions that have low solubility in water was previously discussed. The
new compounds precipitate out of the water and form a protective barrier on the pipe
wall. In contrast, polyphosphate compounds are known to bond with metal ions, bringing
M58 book.indb 17 11/17/2010 4:24:35 PM

them into their molecular structure and holding these metal ions in solution. That bond-
ing feature is the very reason that polyphosphates have traditionally been used in water
distribution systems to “sequester” iron and manganese. That is, iron and manganese ions
in the water are taken up by the added polyphosphates and not allowed to form other
compounds where they might precipitate as solids on water customers’ sinks and laundry.
Many years ago, it was not realized that not only was iron and manganese kept in solution,
but also other metal ions, such as lead and copper, that happened to be in the water was
captured. This polyphosphate/metal mixture is consumed by the water utility customer.
Since the publishing of the LCR in 1991, polyphosphates have been marketed as a
means to both sequester iron and manganese and minimize lead and copper in the system
by precipitation of a fine metallic compound as a barrier on the pipe wall. How can such
compounds possibly perform these two opposing functions? Researchers have found the
answer. It turns out that polyphosphates can perform these two opposing functions but
not at the same time. When the polyphosphate product is first introduced into the distribu-
tion system, it is in its polymeric form and sequesters metals including iron, manganese,
copper, and lead. However, a number of factors in the distribution system environment
break up the polyphosphate compound into orthophosphate ions. These factors include
pH, temperature, time, polyphosphate concentration, polyphosphate-to-orthophosphate
ratio, type of polyphosphate species, and the presence of other various constituents in the
water. When the polyphosphate has reverted back to orthophosphate ions in this way, it
has the passivating effect on the pipe wall described previously (Klueh and Robinson 1988;
Koudleka et al. 1982; Zinder et al. 1984).
A problem with adding polyphosphate as an agent to minimize lead and copper
release is that there is no way to predict where and when the polyphosphate will revert to
orthophosphate. Even if a prediction could be made, there would most likely be some loca-
tions and time frames in which water utility customers were consuming the higher metal
concentrations until the reversion to orthophosphate could take place.
Past technical literature has warned against using polyphosphates (AwwaRF and
TZW 1996). One study showed lead concentration increasing to four times that of untreated
water with the addition of a polyphosphate–orthophosphate blend (Cantor et al. 2000). It
was best summed up in 1991 with the statement, “The application of polyphosphates for
the specific purpose of lead corrosion control entails considerable uncertainty and risk”
(Holm and Schock 1991).
Even so, it is important to keep in mind that polyphosphate compounds and polyphos-
phate–orthophosphate blended products differ from manufacturer to manufacturer and
are typically proprietary formulations. It is possible that some formulations can achieve a
balance between metal sequestration and minimization of metal release. However, more
rigorous process control must be implemented to show that water quality goals are being
achieved in a distribution system. In addition, more rigorous proof of a formulation’s prop-
erties should be required before a formulation is pumped into a water system.
Zinc orthophosphate. The use of zinc orthophosphate in attempts to minimize lead
in drinking water has also caused much confusion. It is theorized that zinc forms car-
bonates that block cathodic sites on the pipe wall while phosphate forms compounds
blocking anodic sites (AwwaRF and TZW 1996). There may be advantages for use of this
chemical for galvanized iron, where the zinc lining is replenished (Grasha et al. 1981), and
for cement-mortar–lined piping, where calcium leaching is inhibited (AwwaRF and TZW
1996). However, there is no evidence that zinc orthophosphate offers more benefit for lead
control with other piping materials than other orthophosphate compounds.
Also, zinc causes environmental problems for wastewater treatment plants in terms
of finished water discharge and sludge disposal. These environmental factors should be
considered when weighing chemical options for corrosion control.
M58 book.indb 18 11/17/2010 4:24:35 PM

Sodium silicate. Sodium silicate is sometimes used as an agent to minimize lead

concentrations in water; however, little research and field information are available to
prove its effectiveness. The limited work that has been done has shown that a relatively
high concentration of silicate is needed (around 20 mg SiO2/L) and that it may take months
to see any reduction in lead concentrations (AwwaRF and TZW 1996). In addition, sodium
silicate is successfully used to sequester iron and manganese—just like the dual use of
polyphosphates. It is not known by what mechanism sodium silicate works both to pre-
vent lead from entering the water and to sequester iron and manganese. It may form a
passivating compound on the pipe wall, or it may adsorb lead, iron, and manganese ions
to the surface of a silicate film on the pipe wall (AwwaRF and TZW 1996). Sodium silicate
has been mostly eliminated from many utilities’ plans of action because of this lack of
information, the increased operating cost associated with the cost of the chemical and
increased dose, and the time it takes to see results. However, because of the rising aware-
ness about the possible negative effects of polyphosphates used in many systems, there
is a renewed interest in the chemical. A 2005 article brought the possibilities of sodium
silicate use back to life by describing both successful reduction of lead and copper and
increased aesthetic qualities in terms of iron and manganese in a water system (Schock
et al. 2005).
Free chlorine and chloramine. Free chlorine, used as a disinfectant in water dis-
tribution systems, is obviously a strong oxidant. One would expect that a strong oxidant
would readily accept electrons at the cathodic sites on the pipe and corrosion would
increase at the anodic sites (AwwaRF and TZW 1996). However, that assumption would
not always be correct. It turns out that under certain conditions where there is little oxi-
dant demand in a system’s water mains, the presence of free chlorine pushes the ORP
of the water high enough to allow lead to lose four electrons instead of two (AwwaRF
and TZW 1996). When lead loses four electrons, lead can form the compound lead diox-
ide (PbO2). This compound is highly insoluble and creates an excellent protective barrier
on the pipe wall against uniform corrosion, thereby lowering the lead concentration in
the water (Lytle and Schock 2005; Cantor et al. 2003). If the addition of free chlorine is
stopped, the lead dioxide layer is lost and lead concentration in the water increases. This
theory is one of several explanations as to what happened in the Washington, D.C., system
in 2003 where the lead concentration increased after a switch of disinfectants from free
chlorine to chloramine, a weaker oxidant. (Refer to the Washington, D.C., case study in
Appendix B for more details.)
Copper
DIC and other naturally occurring ions. The approach to understanding uniform
corrosion in other metals is similar to that used for understanding uniform corrosion in
lead, but each metal has its own special chemistry that must be considered. For instance,
copper is a metal that is hesitant to give up its electrons. Inevitably, drinking water in a
distribution system contains oxygen and disinfectants that are strong enough to pull the
electrons away, but it is harder to pull electrons away from copper than from lead.
The most common forms of copper corrosion by-products are Cu2O (cuprite) and
CuO (tenorite) (AwwaRF and TZW 1996). When the pH is high, these two compounds are
very insoluble and are considered to form a passivating film on the pipe wall (AwwaRF and
TZW 1996). However, various pH, alkalinity, and DIC combinations push these compounds
toward higher solubility. Other copper oxides and hydroxides of high solubility may also
form at various pH, alkalinity, and DIC combinations (AwwaRF and TZW 1996).
Copper has been more difficult to work with than lead in a theoretical solubility
model. There are many unanswered questions about the conditions under which the cor-
rosion by-products form, and there are questions about the thermodynamic solubility
constants associated with the corrosion by-products (Schock et al. 1994). Nevertheless,
M58 book.indb 19 11/17/2010 4:24:36 PM

attempts at modeling copper solubility and observations of field data provide an under-
standing that once again DIC and pH play an important role in the uniform corrosion of
copper.
When the pH is less than 6.0 at any DIC or alkalinity, uniform corrosion of copper
is enhanced because the copper corrosion by-products are highly soluble at acidic and
slightly acidic pH values (AwwaRF and TZW 1996). At moderate alkalinity, bicarbonate
is available to form protective films with the copper (AwwaRF and TZW 1996). However,
at high DIC levels (as expressed as high alkalinity), the corrosion by-products are more
soluble and result in higher metal release (AwwaRF and TZW 1996). Increasing the pH at
higher alkalinity may be helpful in reducing metal release (AwwaRF and TZW 1996). How-
ever, at an alkalinity of about 200 mg/L as CaCO3 or more, pH adjustment is not effective
(Shock et al. 1995).
The preceding observations are only a general view of copper corrosion. It has been
noted in the technical literature that when the pH is greater than 7.0, uniform corrosion
of copper occurs very slowly and that most of the copper corrosion observed is a pitting
(nonuniform) corrosion (AwwaRF and TZW 1996). Pitting corrosion will be discussed later
in this chapter.
Orthophosphate. Orthophosphate has been found to be effective in lowering copper
release into water from uniform corrosion where the pH is above 7.0. At acidic pH val-
ues, orthophosphate is ineffective in lowering copper release (AwwaRF and TZW 1996).
At higher pH values around 8.0 where copper release is diminished because of the pH
increase, orthophosphate addition shows no additional benefit (AwwaRF and TZW 1996;
Schock and Sandvig 2006). These facts point out that first achieving an optimum pH is
important before considering the use of orthophosphate for controlling copper levels.
Polyphosphates and blends. The previous discussion concerning lead pointed out
how the use of polyphosphate and some orthophosphate–polyphosphate blends to miti-
gate uniform corrosion often has the opposite effect. It is the polyphosphates’ reversion to
orthophosphate that gives it the ability to minimize metal release if conditions are favor-
able. Many studies reported in the technical literature have shown that copper levels can
also increase dramatically with the addition of polyphosphates and phosphate blends
(AwwaRF and TZW 1996). Studies carried out on polyphosphate blend addition in hot
water showed a benefit in reduction of copper release, but it was also acknowledged that
the higher temperatures push the polyphosphate to revert to orthophosphate (AwwaRF
and TZW 1996).
One case study by Knobeloch et al. exemplifies the point that the use of polyphos-
phates and blends is risky. Some water customers in one city experienced staining of sinks
and staining of hair and fell ill with gastrointestinal upsets. Their malaise was traced to
high copper levels in the drinking water (Knobeloch et al. 1998). A polyphosphate blend
had recently been added to sequester dissolved iron in the groundwater source. An inves-
tigation proved that by removing the polyphosphate blend from the drinking water, cop-
per levels in new homes were lowered dramatically (Cantor et al. 2000). This case study
is related here as a warning of what can happen when polyphosphates are added to a
drinking water system and are not properly monitored. The literature is clear that metals
will increase in concentration in the presence of polyphosphates until the polyphosphates
revert to orthophosphate.
Sodium silicate. As previously discussed, there is not enough literature to assess
the effectiveness of sodium silicate as a means to provide a protective barrier between
metal and water. In studies that have been performed, it cannot be determined if the
silicate decreases the release of copper into the water or if the increase in pH as a
result of sodium silicate addition gives the benefit of copper control (AwwaRF and
TZW 1996).
M58 book.indb 20 11/17/2010 4:24:36 PM

Free chlorine. Observations have been made that free chlorine may increase the
release of copper into the water, but it does not necessarily increase it enough to push a
water system out of compliance with the LCR (Cantor et al. 2003).
Copper Alloys
Copper alloys are typically present as pipe fittings and solders in water distribution
systems. Brasses, which are alloys of copper and zinc, are prevalent. Brasses, like other
copper alloys, also contain minute quantities of a number of other metals such as lead and
arsenic in order to impart specific physical and chemical characteristics to the alloy.
Copper alloys can release various metals to the water in a number of ways including
dezincification, dissolution, galvanic corrosion, impingement, and pitting. Detailed discus-
sions can be found elsewhere (AwwaRF and TZW 1996).
There is much debate over the quantity of lead that should be allowed in a copper
alloy, over assessing the significance of lead contribution from the dissolution of copper
alloys, and over the method of assessing copper alloys’ potential for leaching lead (Dudi
et al. 2005; Purkiss 2005; Edwards and Schock 2005). However, some completed studies
indicate that lead released from brass alloys is significant in its contribution to the lead
concentration in the drinking water (Boyd et al. 2008).
Other studies may also point to the need to replace lead in the alloys. But care must
be taken not to use other metals where the dissolution characteristics and health effects
are not well understood (Curtis 2006).
Iron
DIC and other naturally occurring ions. As with lead and copper exposed to water,
iron exposed to water is affected by oxygen and other oxidants as electron acceptors in
the electrochemical battery set in motion by uniform corrosion. And once again, alkalinity,
DIC, and pH play a major role in forming compounds with the metal ions that are released
into the water. As with lead and copper, the less soluble the corrosion by-products, the
lower the concentration of iron found in the water.
If the corrosion by-product formation ends with the production of FeCO3 (siderite),
a relatively nonporous scale is formed to stop the migration of oxidants to the cathodic
sites on the pipe wall. But, if excess oxygen is present in the water, the by-product reac-
tions continue and other less dense and more porous iron scales are formed. In this case,
oxygen and ions can migrate to the pipe wall through the scale as well as react with com-
pounds in the scale itself (AwwaRF and TZW 1996).
In general, the corrosion scale for iron pipe is quite complex and is typically described
as having three distinct layers. A summary of the physical and chemical characteristics
of the layers and factors that influence iron release are described elsewhere (Burlingame
et al. 2006).
Orthophosphate. Orthophosphates have been shown to reduce the corrosion rates
of iron and also to stabilize the iron scale that forms on the pipe surface (AwwaRF and
TZW 1996).
Polyphosphates and blends. Polyphosphates and blends have been shown to increase
iron uniform corrosion rates and to destabilize the iron scale on the pipe surface (AwwaRF
and TZW 1996). That is to say, the same chemical mechanism that allows polyphosphate
to be used as a sequestering agent for iron actively pulls iron scale from pipe surfaces and
holds the iron in solution, thereby increasing iron concentration in the water.
Free chlorine. Observations have been made that free chlorine increases the release
of iron into the water (Cantor et al. 2003). In addition, the physical characteristics of the
iron scale depend on the rate of reaction between the iron and the oxidant. Free chlorine
reacts rapidly with soluble iron forming more porous and less dense solids than those
formed in a slower reaction with oxygen or chloramine (Burlingame et al. 2006).
M58 book.indb 21 11/17/2010 4:24:36 PM

Galvanized Iron Pipe

Galvanized iron pipe is a steel pipe with a layer of zinc applied on the surface. In
an aqueous environment, zinc forms a protective surface layer and corrodes more slowly
than iron would in direct contact with the water. When zinc does corrode away from the
iron or when the zinc layer does not adequately cover the iron, then the iron corrodes in
the aqueous environment (AwwaRF and TZW 1996).
With galvanized pipe, consideration must be given to the corrosion by-products of
zinc and other impurities. Zinc corrosion by-products have been observed to form a pro-
tective layer in neutral to slightly alkaline water where the pH is between 7.0 and 8.5. In
terms of uniform corrosion, galvanized pipe has acceptable qualities for most plumbing
systems (AwwaRF and TZW 1996). Zinc orthophosphate has been found to be beneficial
for slowing metal release from galvanized iron pipe (Grasha et al. 1981).
Once the zinc layer has been removed by corrosion, the iron layer is exposed. The
corrosion of iron was previously discussed and is summarized elsewhere (Burlingame
et al. 2006).
Cement-Lined and Similar Pipe
New ductile iron pipe installed in distribution systems is typically cement–mortar
lined, so attention is focused on the integrity of the lining and its impact on water quality.
This discussion also applies to concrete pipe and asbestos–cement pipe. Several aspects
of cement must be considered.
• Metal impurities in the cement mix can leach into the water (Guo et al. 1998; Berend
and Trouwborst 1999).
• Calcium from the cement can dissolve into the water. It is the one circumstance
when calcium carbonate saturation is an important factor in the dissolution of the
pipe material (AwwaRF and TZW 1996). If the cement lining comes in contact with
water that is far away from the calcium carbonate saturation point, the calcium in
the cement can be dissolved into the water. If the cement lining comes in contact
with water that is close to, at, or past the calcium carbonate saturation point, the
cement lining will have a longer life. So, the integrity of cement-lined pipe does
depend on such calculations as the Calcium Carbonate Precipitation Potential and
the Langelier Index.
• Calcium silicate is also a major component of cement materials. The solubility of
calcium silicate has been noted as an important aspect to consider for maintaining
the integrity of cement-lined pipe (Trussel and Morgan 2007).
• Other ions, such as sulfate, in the water in contact with the cement lining can
affect its integrity by forming hydrated compounds that cause the cement to crack
• Zinc in salts of orthophosphate, sulfate, and chloride have been found to slow the
leaching of calcium (AwwaRF and TZW 1996).
• Polyphosphates have been found to accelerate the leaching of calcium (AwwaRF
and TZW 1996).
• Release of calcium compounds from the cement can significantly increase the pH
of the water thereby affecting pH-dependent chemical interactions in the distribu-
tion system, such as the effectiveness of corrosion control chemicals, DBP forma-
tion, and scaling (AwwaRF and TZW 1996).
It is important to note, though, that the integrity of cement-linings has nothing to do
with the electrochemical process of corrosion. The mechanisms of concern are compound
deposition, diffusion, and solubility.
M58 book.indb 22 11/17/2010 4:24:36 PM

Mechanisms of Metal Release by nonuniform Corrosion_

In nonuniform corrosion, the anode and cathode are stationary and a localized loss of
metal occurs at the anode. In some cases, solid corrosion by-products build up in that local-
ized area while a pit forms in the metal pipe underneath. The pit may eventually extend to
the outside of the pipe creating a pinhole leak. In other cases, pitting is not detected while
nonuniform corrosion occurs.
Following are common mechanisms of nonuniform corrosion. With all of these forms
of nonuniform corrosion, pitting and corrosion debris may or may not be observed.
Pipe Manufacturing Quality

The manufacturing quality of a pipe can determine the fate of the pipe in an aqueous envi-
ronment. The more homogeneous a pipe’s structure, the less likely that localized anodic
and cathodic sites will develop. It has been found that in copper pipe, for instance, hetero-
geneity of copper crystal structure can cause pitting to occur when the pipe is in service
Also, residual films, such as carbon films used as lubricant on copper pipe in the
manufacturing process, can set up the conditions for pitting (AwwaRF and TZW 1996).
Pipe Installation Practices

Any irregularities on a pipe’s surface may set up the conditions for pitting. Spillage of sol-
dering fluxes onto the interior pipe wall is one source of irregularities. Incorrect joints and
bends may also cause irregularities.
Galvanic Corrosion
Metals have individual potentials for giving up electrons. A metal that is more willing to
give up an electron is said to be “lower in the galvanic series of metals” than other metals
with less of a tendency. If two dissimilar metals are connected, a flow of electrons can
begin because the components of a battery have been assembled. That is, the metal lower
in the galvanic series acts as the anode where electrons are given up. The metal higher in
the galvanic series acts as the cathode. The electrons easily flow through both metals as
the medium (wire) connecting the anode to the cathode.
In the galvanic series of metals, zinc has more of a tendency to give up electrons than
iron. Iron has more of a tendency than lead; lead more than brasses and copper.
For example, if a large area of brass is connected to a galvanized iron pipe (zinc-
coated), then galvanic corrosion will occur, causing pitting of the zinc coating and ulti-
mately the iron pipe underneath (AwwaRF and TZW 1996).
In some cases, plumbers install a dielectric union between two dissimilar metals,
which inserts a nonconducting material between the two metals, acting as an off-switch
would in an electric circuit.
Concentration Cells
Heterogeneity can occur not only in the structure of metal pipe but also in the water in
contact with the pipe. For instance, if there is a pocket of higher oxygen concentration in
one location versus another, then a corrosion potential can develop (AwwaRF and TZW
1996). For example, low dissolved oxygen in water may be captured between a bolt and
a metal surface that is still in contact with higher dissolved oxygen in the general water
solution in the main pipe or tank. Or low dissolved oxygen may occur under pipeline debris
in contrast with the oxygenated solution outside of the debris (AwwaRF and TZW 1996).
M58 book.indb 23 11/17/2010 4:24:36 PM

Erosion Corrosion
Water at high velocity may physically scour pipe walls, causing surface irregularities and
encouraging corrosion to occur locally (AwwaRF and TZW 1996). This scouring typically
occurs at pipe bends in systems where velocity and possibly temperature are factors. A
characteristic horseshoe-shaped pit is formed, typically just above a pipe bend, in the
direction of flow (Figure 2-4).
Chloride and Sulfate

In 1994, researchers found that the presence of chloride tended to decrease the likelihood
of copper pitting. Even though chloride is aggressive toward copper, it forms a passivating
scale that prevents further corrosion (Edwards et al. 1994). Sulfate was found to initi-
ate and propagate pitting and a low chloride-to-sulfate ratio encourages copper pitting
(Edwards et al. 1994). However, in 2007 researchers found that a low chloride-to-sulfate
ratio is beneficial for preventing galvanic corrosion, especially to decrease lead leaching
from brass (Triantafyllidou and Edwards 2006).
Other Influencing Chemicals

The presence of natural organic matter (NOM) appears to inhibit pitting, and when NOM is
removed by water treatment, pitting increases (Edwards et al. 1994). However, this obser-
vation may be confounded with the possible increase in assimilable organic carbon (AOC)
as NOM is removed, with subsequent possible pitting initiated by microorganisms using the
carbon for growth (Escobar and Randall 2002; Cantor et al. 2006). Another reported case
study describing failure of copper water service lines just after construction of new homes
pinpointed carbon dioxide as the cause of the pitting corrosion (Cohen and Meyers 1987).
Stray Electrical Currents and Electrical Grounding

Stray electrical currents and electrical grounding may play a role in metal loss from
pipes and pitting of pipes. However, the damage that occurs is found on the exterior of
the pipe where the electrical current leaves the pipe and flows into the soil (Duranceau
et al. 1998).
This concept is confusing. Some brochures and Web sites seek to explain lead and
copper corrosion to the layman but do not differentiate between the external pitting phe-
nomenon and internal pitting with increased metal concentration in the water. General
statements are made, such as: “Corrosion of pipes is greater if grounding wires are con-
nected to them” (Kissel et al. 2003). This author found no references to show that stray
currents or grounding increase metal concentration in water or increase internal pitting.
Stray electrical currents and grounding currents should, nevertheless, be taken into
consideration in a corrosion investigation. If electrical currents are suspected as a possi-
ble factor, measurements must be taken to prove or disprove the theory. The phenomenon
and measurement techniques are described in an article by Duranceau and colleagues
(1998) and in publications by the National Association of Corrosion Engineers (NACE).
Other mechanisms of nonuniform corrosion must be ruled out as well.
Microbially Influenced Corrosion

A greatly overlooked cause of pitting corrosion is microbially influenced corrosion
(MIC). A common question regarding MIC is, “how can the microorganisms get nourish-
ment from and thrive in the toxic environment of the metal?” The answer is that the micro-
organisms are not getting nourishment from the pipe. They are getting nourishment from
nutrients in the water. They can thrive in the toxic environment of the metal, such as
copper, because they secrete an enzyme to protect themselves from their environment.
M58 book.indb 24 11/17/2010 4:24:36 PM

Source: Photo courtesy of Process Research Solutions LLC.
Figure 2-4 Characteristic horseshoe-shaped pit of erosion corrosion
That enzyme is acidic. The localized drop in pH helps to initiate the corrosion at that site
(Bremer et al. 2001). One study further elucidates the influence that microorganisms most
likely have on pitting corrosion. Bacteria were injected into artificial tap water to study
the pitting of copper and brass. It was found that the pitting occurrence was measurably
larger in the presence of bacteria (Valcarce et al. 2005). Iron pipe may also fall victim
to MIC. There is broad acceptance that iron pipe and galvanized pipe can form pits and
tubercles. It has been found that sulfate-reducing bacteria (SRB) play an important role
within iron tubercles, changing their chemical and physical properties (Burlingame et al.
2006; Lytle et al. 2005).
In looking for telltale signs of MIC, start with piping configuration. MIC typically
is initiated in a system at the farthest reaches of the distribution system, at dead ends,
and at the low points of pipelines. These are the points where water stagnates, sediment
accumulates, and disinfection levels are typically low (Bremer et al. 2001). For example, if
all pipe failures are occurring underneath the concrete slab foundation of buildings, then
suspect MIC. If a problem is not investigated at this early stage, however, the corrosion can
“spread” to other locations just as an infection spreads in a person and the original loca-
tion of the problem may be obscured (Cantor et al. 2003; Cantor et al. 2006).
Next, the presence or absence of microorganisms can be identified. In water samples,
heterotrophic plate count (HPC) analysis using R2A media may be performed to check for
the relative presence of heterotrophic microorganisms. (This and other water analyses
will be described in chapter 3.) HPC is also an indicator that other types of microorgan-
isms may be thriving at the same location.
If corroded pipe samples can be obtained, the corrosion debris may be analyzed for
the presence of microorganisms using staining and microscopic techniques (Standridge
et al. 2003).
Another important aspect of assessing the possibility of MIC is to assess the biosta-
bility of the water. The term biostability is used to describe the waterborne nutrients that
can encourage and support the growth of microorganisms in the distribution system as
balanced with disinfection to prevent the growth of microorganisms. Excellent literature
on biostability is available (Van der Kooij 1992; Volk and LeChevallier 2000; Escobar and
Randall 2002; Zhang et al. 2002). However, biostability has not been associated directly
with controlling metal release in a distribution system or with piping system integrity.
Experience with pitting corrosion emphasizes how essential biostability is (Cantor et
al. 2006). In terms of biostability, there is a good body of literature and awareness con-
cerning nitrification in a distribution system, the result of microbiological activity where
excess ammonia-nitrogen is present (Fleming et al. 2005; AWWA 2006). However, other
M58 book.indb 25 11/17/2010 4:24:38 PM

potential sources of microbiological nutrients, such as phosphorus, sulfate, and carbon,

should also be identified if present (Srinivasan and Harrington 2007). There should also
be awareness that some disinfectants are more powerful than others. While chloramine
disinfection has advantages over free chlorine as a disinfectant in some water systems,
a higher dose of chloramine is required to equal the protective level of free chlorine at a
lower dose. For instance, it would not be uncommon for a water system using chloramine
to use a dose of 2.0 to 2.5 mg/L total chlorine (Zhang et al. 2002) as opposed to 1 mg/L or
less of free chlorine.
Another aspect involved in pinpointing MIC in a system is to study the relative con-
centration of metals at sites around the distribution system. Many times a water system
is in compliance with the LCR, meaning that 90 percent of the lead results and 90 percent
of the copper results do not exceed the action levels (ALs) for lead or copper. However,
there may be some sites where the metal levels are an order of magnitude higher than
other sites. One cannot help but wonder why the same water characteristics would cause
greater metal release at some sites and not at others. A comparison of HPC levels and
metal levels at these sites may point to a high number of microorganisms being present
when the metal level is high, signaling that MIC could be occurring in the water system.
Source: Photo courtesy of Process Research Solutions LLC by permission of NACE.
Figure 2-5 Example of microbially influenced pitting of copper pipe
Source: Photo courtesy of Process Research Solutions LLC.
Figure 2-6 Another example of microbially influenced pitting of copper pipe
M58 book.indb 26 11/17/2010 4:24:44 PM

A common view of microbially influenced pitted pipe is shown in Figure 2-5. Here
the corrosion by-products from the pitting corrosion build up in mounds or tubercles over
the pits that have been formed. Therefore, sampling the water for the particulate form
of a metal can indicate where the pitting is occurring (Cantor et al. 2003; Cantor 2006).
However, the lack of particulate copper does not mean that a pipe is not pitted or that
nonuniform corrosion is not occurring. Some cases of microbiologically influenced pitting
corrosion do not develop numerous pits as shown in Figure 2-6. Just like the other forms of
nonuniform corrosion, the problem can initially appear to be uniform corrosion.
Further research concerning the nuances of MIC is greatly needed so that aware-
ness of the role it plays in metal release may be increased. In the meantime, investigators
should always consider the role of microorganisms in a given problem whether it is a pit-
ting corrosion question, a uniform corrosion question, or a significant difference in metal
levels seen at some sites versus others.
Other Mechanisms of Metal Release_____________________

Adsorption and Release by Chemical Scales
Another greatly overlooked cause of metal release in a water distribution system is the
adsorption of metals, such as lead and copper, by chemical scales in the pipelines, such as
iron, manganese, and aluminum, which have precipitated in the distribution system.
Recall that hydrous manganese oxide is used to adsorb and remove radium as a water
treatment process. Iron-oxide–coated sand or coagulation with iron is used to adsorb and
remove arsenic as a water treatment process. Aluminum salts are found in soils and are
known to sorb other metals. The focus needs to be on the fact that these scales that build
up in water distribution system pipelines have the same adsorptive properties.
Studies from one water system, the source water of which includes dissolved iron
and manganese from the aquifer, showed that iron and manganese scales that precipitate
in the distribution system adsorb and accumulate various metals. Pipe film analyses from
the water system showed the presence of iron and manganese scales with adsorbed lead
(Schock et al. 2006; Maynard and Mast 2006). This finding corresponded with lead moni-
toring studies that had shown lead occurring at the consumers’ taps at random times and
predominantly in particulate form (Cantor 2006).
Iron and manganese are not the only chemical scales that have an affinity for metal.
Aluminum compounds have been studied in other contexts as being “metal scavengers”
(Schock 2005).
In summary, it appears that the water contains minute quantities of various dissolved
metals from contact with equipment in the system, trace metals in treatment chemicals,
and trace metals in water sources. Those minute quantities are adsorbed by scales in the
distribution system. If the scales are not cleaned from the pipelines efficiently, the minute
quantities of metals accumulate. The accumulated metals may sometimes desorb back
into the water. Or, when particles of the scales break off and flow to the consumers’ taps,
they carry the other metals with them.
System Maintenance
There have also been water systems where the presence of high lead concentrations, spe-
cifically particulate lead, has been related to oversights in water system maintenance. In
one case, the lining of a building’s water tank had deteriorated and formed sludge on the
bottom of the tank with debris carried over to the building’s water system periodically.
The lining contained lead that was carried to the drinking water taps. Gaskets in the sys-
tem had also deteriorated and lead impurities were carried into the water (Cantor 2006).
Therefore, the integrity of tank linings and gaskets may impact water quality.
M58 book.indb 27 11/17/2010 4:24:44 PM

Summary_________________________________________________
This chapter has described the common mechanisms by which metals are released into
the water in a distribution system. It is obvious that the actual electrochemical process
that is termed corrosion and typically thought of as uniform corrosion is only one of many
aspects involved in metal release.
When achieving “optimal corrosion control is discussed,” as stated in the LCR, the
desire is to keep metal concentrations in the water as low as economically feasible. Per-
haps it should be expressed as “achieving the least possible metal release.” The only way
to achieve that goal is to determine and then specifically to address the offending mecha-
nism or mechanisms that are increasing the metal concentrations.
Additional Reading______________________________________
Much of the knowledge that we have today on control of internal corrosion and metal
release in water distribution systems is summarized in the book Internal Corrosion of
Water Distribution Systems, published by the AwwaRF and its German counterpart, TZW
(TZW) (second edition, 1996). The book is a history and literature review of the corrosion
topic into the mid-1990s. The elements of understanding metal release in drinking water
are all there, described by the insightful researchers of the past.
References________________________________________________
American Water Works Association (AWWA). Cantor, A.F., J.B. Bushman, M.S. Glodoski, E.
2006. M56: Fundamentals and Control of Kiefer, R. Bersch, and H. Wallenkamp.
Nitrification in Chloraminated Drink- 2006. Copper Pipe Failure by Microbio-
ing Water Distribution Systems. 1st ed. logically Influenced Corrosion. Materials
Denver, Colo.: AWWA. Performance, 45(6):38.
American Water Works Association Research Cantor, A.F., D. Denig-Chakroff, R.R. Vela, M.G.
Foundation and DVGW-Technologiezen- Oleinik, and D.L. Lynch. 2000. Use of Poly-
trum Wasser (AwwaRF and TZW). 1996. phosphate in Corrosion Control. Jour.
Internal Corrosion of Water Distribution AWWA, 92(2):95.
Systems, 2nd ed. Denver, Colo.: AwwaRF Cantor, A.F., J.K. Park, and P. Vaiyavatjamai.
and AWWA. 2003. Effect of Chlorine on Corrosion in
Berend, K., and T. Trouwborst. 1999. Cement– Drinking Water Systems. Jour. AWWA,
Mortar Pipes as a Source of Aluminum. 95(5):112.
Jour. AWWA, 91(7):99. Cohen, A., and J.R. Meyers. 1987. Mitigating
Boyd, G.R., G.L. Pierson, G.J. Kirmeyer, and R.J. Copper Pitting Through Water Treatment.
English. 2008. Lead Variability Testing Jour. AWWA, 79(2):58.
in Seattle Public Schools. Jour. AWWA, Colling, J.H., P.A.E. Whincup, and C.R. Hayes.
100(2):53–64. 1987. The Measurement of Plumbosol-
Bremer, P.J., B.J. Webster, and D.B. Wells. 2001. vency Propensity to Guide the Control of
Biocorrosion of Copper in Potable Water. Lead in Tapwaters. Jour. Inst. Water and
Jour. AWWA, 93(8):82. Environmental Management, 1(3):263.
Burlingame, G.A., D.A. Lytle, and V.L. Snoeyink. Curtis, T.W. 2006. Letter from AWWA to USEPA
2006. Why Red Water? Understand Iron regarding National Primary Drinking
Release in Distribution Systems. Opflow, Water Regulations for Lead and Copper:
32(12):12. Short-Term Regulatory Revisions and
Cantor, A.F. 2006. Diagnosing Corrosion Prob- Clarifications (Docket ID No. EPA-HQ-
lems Through Differentiation of Metal OW-2005-0034) (Sept. 18, 2006).
Fractions. Jour. AWWA, 98(1):117. Dudi, A., M. Schock, N. Murray, and M. Edwards.
Cantor, A.F., J.B. Bushman, and M.S. Glodoski. 2005. Lead Leaching From Inline Brass
2003. A New Awareness of Copper Pipe Devices. Jour. AWWA, 97(8):66.
Failures in Water Distribution Systems. In Duranceau, S.J., M.J. Schiff, and G.E.C. Bell.
Proc. of the AWWA Water Quality Tech- 1998. Electrical Grounding, Pipe Integrity,
nology Conference. Denver, Colo.: AWWA. and Shock Hazard. Jour. AWWA, 90(7):40.
M58 book.indb 28 11/17/2010 4:24:44 PM

Edwards, M., J.F. Ferguson, and S.H. Reiber. Purkiss, D. 2005. We’ve Got Mail. Jour. AWWA,
1994. The Pitting Corrosion of Copper. 97(11):8.
Jour. AWWA, 86(7):74. Schock, M.R. 1980. Response of Lead Solubility
Edwards, M., and M.R. Schock. 2005. We’ve Got to Dissolved Carbonate in Drinking Water.
Mail. Jour. AWWA, 97(11):10. Jour. AWWA, 72(12):695.
Escobar, I.C., and A.A. Randall. 2002. Case Schock, M.R. 2005. Distribution Systems as
Study: Ozonation and Distribution System Reservoirs and Reactors for Inorganic
Biostability. Jour. AWWA, 93(10):77. Contaminants. In Distribution Sys-
Fleming, K.K., G.W. Harrington, and D.R. Nogu- tem Water Quality Challenges in the
era. 2005. Nitrification Potential Curves: A 21st Century—A Strategic Guide. M.J.
New Strategy for Nitrification Prevention. McPhee, ed. Denver, Colo.: AwwaRF.
Jour. AWWA, 97(8):90.30 Schock, M.R., T.L. Geske, M.K. DeSantis, and
Grasha, L.A., W.W. Risner, and M.K. Davis. 1981. R.C. Copeland. 2006. Scale Analysis for
Update on Corrosion Study at the Metro- Madison, WI Samples. Private report
politan Water District of Southern Califor- from USEPA (Apr. 6, 2006).
nia. AWWA Research Foundation Water Schock, M.R., D.A. Lytle, and J.A. Clement. 1994.
Quality Research News (September 1981). Modeling Issues of Copper Solubility in
24(9):2–3. Drinking Water. In Proc. of the National
Guo, Q., P.J. Toomuluri, and J.O. Eckert Jr. 1998. Conference on Environmental Engineer-
Leachability of Regulated Metals From ing, Critical Issues in Water and Waste-
Cement–Mortar Linings. Jour. AWWA, water Treatment. Boulder, Colo.: ASCE.
90(3):62. Schock, M.R., D.A. Lytle, and J.A. Clement.
Holm, T.R., and M.R. Schock. 1991. Potential 1995. Effect of pH, DIC, Orthophosphate
Effects of Polyphosphate Products on and Sulfate on Drinking Water Cupro-
Lead Solubility in Plumbing Systems. solvency. EPA/600/R-95/085. Cincinnati,
Jour. AWWA, 83(7):76. Ohio: USEPA, Office of Research and
Kissel, D.E., P.R. Vendrell, and J.H. Atiles. 2003. Development. http:www.epa.gov/ORD/
Your Household Water Quality: Lead and WebPubs/effect.
Copper. Consumer information brochure Schock, M.R., D.A. Lytle, A.M. Sandvig, J. Clem-
from the University of Georgia Coopera- ent, and S.M. Harmon. 2005. Replacing
tive Extension Services. Polyphosphate With Silicate to Solve Lead,
Klueh, K.G., and R.B. Robinson. 1988. Sequestra- Copper and Source Water Iron Problems.
tion of Iron in Groundwater by Polyphos- Jour. AWWA, 97(11):84.
phates. Jour. Env. Engr., 114(5):1192. Schock, M.R., and A.M. Sandvig. 2006. Long-
Knobeloch, L., C. Schubert, J. Hayes, J. Clark, Term Impacts of Orthophosphate Treat-
C. Fitzgerald, and A. Fraundorff. 1998. ment on Copper Levels. In Proc. of the
Gastrointestinal Upsets and New Copper AWWA Annual Conference and Exposi-
Plumbing—Is There a Connection? Wis- tion. Denver, Colo.: AWWA.
consin Medical Jour. (Jan.). Srinivasan, S., and G. Harrington. 2007. Bio-
Koudleka, M., J. Sanchez, and J. Augustynski. stability Analysis for Drinking Water
1982. On the Nature of Surface Films Distribution Systems. Wat. Research,
Formed on Iron in Aggressive and Inhibit- 41(4):2127.
ing Polyphosphate Solutions. Jour. Elec- Standridge, J., B. Hoffman, and L. Peterson.
trochem. Soc., 129(6):1186. 2003. Brown Deer Pipe Corrosion Study.
Lytle, D.A., T.L. Gerke, and J.B. Maynard. 2005. Madison, Wis.: Wisconsin State Labora-
Effect of Bacterial Sulfate Reduction tory of Hygiene.
on Iron Corrosion Scales. Jour. AWWA, Triantafyllidou, S., and M.A. Edwards. 2006.
97(10):109. Effect of Coagulant Selection on Lead
Lytle, D.A., and M.R. Schock. 2005. Formation Leaching: Importance of the Chloride to
of Pb(IV) Oxides in Chlorinated Water. Sulfate Mass Ratio. In Proc. of the AWWA
Jour. AWWA, 97(11):102. Water Quality Technology Conference.
Maynard, B., and D. Mast. 2006. Composition of Denver, Colo.: AWWA.
Interior Scales on Lead Source Materials.
In Proc. of the AWWA Water Quality Tech-
nology Conference. Denver, Colo.: AWWA.
M58 book.indb 29 11/17/2010 4:24:44 PM

Trussell, R.R., and J.J. Morgan. 2007. A Satura- Volk, C.J., and M.W. LeChevallier. 2000. Assess-
tion Index for Cement Surfaces Exposed ing Biodegradable Organic Matter. Jour.
to Water. In Proc. of the AWWA Water AWWA, 92(5):64.
Quality Technology Conference. Denver, Zhang, M., M.J. Semmens, D. Schuler, and R.M.
Colo.: AWWA. Hozalski. 2002. Biostability and Microbio-
Valcarce, M.B., S.R. Sanchez, and M. Vazquez. logical Quality in a Chloraminated Distri-
2005. Localized Attack of Copper and bution System. Jour. AWWA, 94(9):112.
Brass in Tap Water: The Effect of Zinder, B., J. Hertz, and H.R. Oswald. 1984.
Pseudomonas. Corrosion Science, 47:795. Kinetic Studies on the Hydrolysis of
Van der Kooij, D. 1992. Assimilable Organic Car- Sodium Tripolyphosphate in Sterile
bon as an Indicator of Bacterial Regrowth. Solution. Wat. Research, 18(5):509.
Jour. AWWA, 84(2):57.
M58 book.indb 30 11/17/2010 4:24:44 PM

AWWA MANUAL M58
Chapter 3
Water Quality Monitoring

and Assessment of
Internal Corrosion
and Increased Metals
Concentrations
Christopher P. Hill
Malcolm Pirnie Inc.
Richard Giani
DC Water
INTRODUCTION_____________________________________________
This chapter elaborates on the fundamentals of internal corrosion outlined in chapter 2. It
emphasizes the critical water quality parameters that affect internal corrosion, discusses
the basis of water quality monitoring programs focused on controlling corrosion, and pres-
ents a general approach to assessing the cause of corrosion-related events (Table 3-1).
31
M58 book.indb 31 11/17/2010 4:24:44 PM


Water Quality Impacts on Corrosion
• Understanding the carbonate balance and its impact on corrosion is critical in that it affects passivation, which
is the ability of water to stop the electrochemical process of corrosion.
• pH is a key component of an effective corrosion control program. Metal dissociation is less likely to occur at
higher pH values (above neutral or higher, depending on water chemistry).
• Maintaining a consistent pH throughout the distribution system is critical to maintaining effective corrosion
control.
• pH, alkalinity, and dissolved inorganic carbon (DIC) are parameters that define the carbonate balance.
• Buffering intensity from carbonate species is highest at pH 6.3 and above 9.0, and lowest in the range 8.0 to 8.5
• Oxidants, and more specifically the residual disinfectant in the distribution system, impact the finished water
oxidation-reduction potential (ORP). Changes in the disinfectant that change the ORP have the potential to
significantly impact corrosion and metals release in the distribution system and home plumbing.
• Anions such as chloride, sulfide, and sulfate can change the corrosive characteristics of the water.
• The presence of sulfide may significantly increase corrosion rates and the occurrence of pinhole leaks in copper
plumbing.
• Chloride has been demonstrated to be a critical anion in copper corrosion and pitting.
• The chloride-to-sulfate mass ratio has been observed to have a significant impact on the leaching of lead from
soldered joints and brass.
• Microbially influenced corrosion (MIC) is often overlooked as a cause of increased metals release. Maintaining
biostability is essential to an effective corrosion control program.
• Corrosion “inhibitors” can be effective for controlling corrosion and metals release; however, it is critical that
the mechanism by which an inhibitor works is understood.
• Water quality may not always be the cause of increased metals release. For instance, hydraulic and other
factors may result in increased metals concentrations.
Water Quality Monitoring Programs
• Lead and Copper Rule (LCR) compliance monitoring may not be sufficient for determining and maintaining
effective corrosion control treatment. More frequent monitoring of LCR and additional parameters is
recommended.
• It is necessary to establish a water quality baseline when developing a corrosion control program to evaluate
impacts of changes in water quality on corrosion.
• A monitoring program should include routine monitoring at the source, entry to the distribution system, and
various locations in the distribution system.
• The number of monitoring locations, the frequency of monitoring, and the number of parameters monitored will
be dependent on sourcewater quality, changes in treatment procedures, and economic factors.
Assessment of Corrosion-related Water Quality Problems
• A convenient way to classify corrosion for the purpose of diagnosis is (1) chemically influenced, (2) microbially
influenced, and (3) physically influenced.
• Identifying the cause of a corrosion-related problem requires that one understands how the three different
categories of corrosion affect that problem.
• Lack of an adequate monitoring program will make it difficult to determine the cause of corrosion-related water
quality problems.
WATER QUALITY CONSIDERATIONS_ __________________________

The Carbonate Balance
Chapter 2 mentioned that the nature of various metallic carbonate species influence a pro-
tective coating on the interior of distribution system piping. Therefore, a thorough under-
standing of the factors that influence the presence of carbonate and bicarbonate species
in water, called the carbonate balance, is necessary in developing an effective corrosion
control program.
M58 book.indb 32 11/17/2010 4:24:45 PM

Water Quality Monitoring and Assessment of Internal Corrosion 33
The pH value. The pH of water is a measure of the strength of the hydrogen ion con-
centration (H+ or H3O+). The pH scale ranges from 0 to 14—values less than 7.0 are consid-
ered acidic, values greater than 7.0 are considered alkaline or basic, and 7.0 is considered
neutral. Drinking water pH values typically range from 6.0 to 10. At higher pH values, there
is less of a tendency for metal surfaces in contact with drinking water to dissolve and dis-
sociate, which is why pH adjustment is a common component of an effective corrosion
control treatment strategy. Maintaining a consistent target pH throughout the distribution
system is always critical to minimizing lead and copper levels at the tap and minimizing
red water occurrence, even if other corrosion control methods are employed.
Another important consideration with regard to pH is its impact on other water qual-
ity parameters discussed in this chapter. It plays a significant role in the carbonate bal-
ance in that it impacts buffer capacity and DIC concentrations. However, it also influences
other corrosion-related parameters, such as ORP and corrosion inhibitor effectiveness.
These impacts are discussed at greater length throughout this chapter.
Alkalinity. Alkalinity is the capacity of water to neutralize acid. It is the sum of car-
bonate (CO32–), bicarbonate (HCO3 –), and hydroxide (OH–) anions and is typically reported
as mg/L as calcium carbonate (mg/L as CaCO3). Waters with high alkalinities tend to have
high buffering capacities, or a strong ability to resist changes in pH. Low alkalinity waters
are less able to neutralize acids or resist changes in pH.
Buffer intensity or buffer capacity. Buffer intensity is a measure of the resistance
of water to upward or downward changes in pH and is a function of pH and alkalinity.
Bicarbonate and carbonate ions are the most important buffering species in most drinking
water supplies. At pH ≥9.0, silicate ions also provide some buffering. Buffering intensity
from carbonate species is normally greatest at approximately pH 6.3 and above 9.0 and
lowest in the range of pH 8.0 to 8.5 (see Figure 3-1).
Buffer intensity can have a significant impact on calcium carbonate precipitation.
Stumm (1960) demonstrated that hard waters with high buffer intensities and slightly posi-
tive Langelier Index (LI) values deposited less calcium carbonate than soft waters with
slightly negative LI. Similar results have been observed by others (Sontheimer et al. 1979).
It was also found that corrosion scales with more than 5 to 10 percent CaCO3 (by weight)
may be less protective than those with lower CaCO3 concentrations (Snoeyink and Wagner
1996). These findings add to the knowledge that calcium carbonate scales should not be
used for corrosion control.
Corrosion reactions require an anode (electron donor) and a cathode (electron recep-
tor). In drinking water distribution systems, these reactions at the metal surface result
in pH shifts at the anode and cathode that further contribute to the corrosion reaction
(Snoeyink and Wagner 1996). Stumm (1960) showed that iron corrosion rates decreased in
a uniform manner as the buffer intensity increased. It was theorized that this pattern was
due to the fact that waters with higher buffer intensity experienced less of a pH increase
at the cathode and less of a pH decrease at the anode. Other studies (Clement and Schock
1998; Van Der Merwe 1988; Pisigan and Singley 1987) have found similar results.
Total dissolved solids/ionic strength. Total dissolved solids (TDS) can also have
a significant impact on corrosion. High TDS concentrations are generally associated with
high concentrations of ions (e.g., Na+, Ca 2+, Mg2+, Cl–, CO32–, SO42–) that increase the con-
ductivity of the water. As discussed earlier, corrosion is an electrochemical reaction in
which electrons from the anodic surface are transferred to the cathodic surface. The
increased conductivity resulting from high TDS concentrations increases the ability of the
water to complete the electrochemical circuit and conduct a corrosive current.
If sulfate and chloride are major anionic contributors to the TDS, the TDS is likely to
show increased corrosivity toward iron-based materials. If the TDS is composed primarily
of bicarbonate and hardness ions, the water may not be corrosive toward iron-based or
cementitious materials but may be highly corrosive toward copper (Schock 1999).
M58 book.indb 33 11/17/2010 4:24:45 PM

Source: Snoeyink and Wagner 1996.
Figure 3-1 Variation of buffer intensity with pH
Low TDS waters may also be corrosive and increase lead solubility. Low TDS waters
often have a strong tendency to dissolve (corrode) materials with which they are in contact
in an attempt to reach electroneutrality. Recall from chapter 2 that uniform corrosion is an
electrochemical process in which the water solution in contact with the cathode provides
the chemicals to accept the electrons donated by the pipe wall. The pipe wall, acting as an
anode, then releases the oxidized metal ion to the water. Water has a limited capacity to
accept dissolved species. Thus, although high TDS waters have many electron receptors,
low TDS waters have the ability to accept a large number of anions and cations, resulting
in a subsequent dissolution of existing pipe scales and corrosion of pipe surfaces.
TDS is a surrogate for the ionic strength of a solution. The ionic strength is a measure
of the strength of the electrostatic field caused by the presence of ions in a solution. More
simply, the presence of anions and cations in solution increases the conductivity of the
M58 book.indb 34 11/17/2010 4:24:45 PM

solution and can increase corrosion unless offset by passivating layers on the pipe surface
(Schock 1999). Ionic strength can be determined as follows (Economic and Engineering
Services and Illinois State Water Survey 1990):
I = 2.5 × 10−5 × TDS
Where:
I = ionic strength
TDS = total dissolved solids concentration, in mg/L
Dissolved inorganic carbon. DIC is the sum of all dissolved inorganic carbon–
containing species and is one of the most critical parameters to controlling internal cor-
rosion. It includes dissolved aqueous carbon dioxide gas (CO2 or H2CO3), bicarbonate ion
(HCO3 –), and carbonate ion (CO32–) in a particular water, and DIC is usually expressed as
milligrams of carbon per liter (mg/L as C) or milligrams of calcium carbonate per liter
(mg/L as CaCO3). Although DIC and alkalinity are similar, they are not the same water
quality parameter.
DIC varies according to water temperature, pH, ionic strength, and alkalinity. An
example of this relationship is provided in Figure 3-2.
DIC also significantly impacts the buffer intensity of water. Figure 3-3 shows that as
the DIC concentration increases, the buffer capacity of the water also increases. Because
DIC controls the buffer capacity in most water systems, sufficient DIC is required to main-
tain a stable pH throughout the distribution system for control of copper and lead (Schock
and Lytle 1995).
DIC, mg C/L
Alkalinity, mg CaCO3 /L
pH=8.5 + pH=9 ◊ pH=9.5 ∆ pH=10 × pH=10.5
Source: Economic and Engineering Services and Illinois State Water Survey 1990.
Figure 3-2 Relationship between alkalinity and DIC for various pH levels (pH = 6.0 to 8.0, I = 0.005,
T = 25°C)
M58 book.indb 35 11/17/2010 4:24:46 PM

Source: Schock 1999.
Figure 3-3 Effect of DIC (measured as C) on buffer intensity
In the presence of bicarbonate, copper corrosion undergoes a critical transition

somewhere between pH 7.0 and 8.5 (Edwards et al. 1994). Scale formed at pH 7.0 catalyzes
oxygen reduction and increases the overall corrosion rate. At pH between 7.5 and 8.5, DIC
complexes dominate copper speciation resulting in increased copper solubility. This effect
is most prominent in the DIC range of 1 to 20 mg C/L. Finally, at pH 8.5 and above, oxygen
reduction rates are unchanged, and copper dissolution is inhibited resulting in a passiva-
tion of the copper surface.
Hardness. Hardness is a characteristic that primarily represents the presence of
dissolved calcium and magnesium in water and is reported as an equivalent quantity of
calcium carbonate (CaCO3). When sufficient calcium and alkalinity are present in waters
with pH greater than the saturation pH, a scale of CaCO3 may form on the inside of distri-
bution piping. However, the formation of carbonate scale can interfere with corrosion con-
trol when other methods, such as phosphate passivation, are employed. It has also been
reported that significant carbonate scales do not form on lead, galvanized, or copper cold-
water pipes, so carbonate scales may not offer the level of protection anticipated (Schock
1999). Therefore, calcium carbonate precipitation is not considered an effective form of
corrosion control, as was once believed. In fact, is it probable in systems that are utilizing
“carbonate precipitation” as a corrosion control method that the actual inhibition of corro-
sion is a result of high pH, alkalinity, DIC, and other factors but that inhibition of corrosion
is not a result of a protective scale of CaCO3 forming on the interior of pipe walls.
Hardness must be taken into consideration when corrosion control is selected and
implemented because it can create scaling problems within the treatment plant and dis-
tribution system infrastructure. In this regard, hardness is an important parameter to be
considered in developing a corrosion control program. However, it is not a stand-alone
indicator of the corrosive nature of a water. As a component of utility corrosion control
programs, many utilities set the finished water pH as high as possible without considering
the potential for scaling.
Corrosivity
Corrosion is the deterioration of a substance or its properties because of an electrochemi-
cal reaction with its environment (Schock 1999). Although this term may occasionally be
used to describe the deterioration of cement–mortar lining and other cement-based mate-
rials used in piping systems, in the context of this manual corrosivity refers to the oxida-
tion and subsequent dissolution of the metal surfaces of the interior of pipe walls. From a
M58 book.indb 36 11/17/2010 4:24:47 PM

regulatory perspective, the concern with lead and copper corrosion is the dissolution, or
increased solubility, of lead and copper. From an aesthetic perspective, the oxidation of
ferrous iron to ferric iron has the potential to create red water and other problems.
The most significant water quality parameters with regard to the corrosivity of water
are pH, alkalinity, and DIC. Balancing pH, alkalinity, and DIC concentrations enables the
formation of carbonate passivating layers. Similarly, pH values less than prescribed corro-
sion inhibitor ranges cause a decrease in chemical passivation and thus cause an increase
in metal solubility.
Edwards and colleagues (1994) demonstrated that increasing the pH from 5.5 to 7.0
reduced corrosion rates by half in new (unaged) copper pipe. Stone et al. (1987) demon-
strated that increasing the pH from 6.0 to 8.0 reduced the copper corrosion rate by half in
both new and aged copper pipe; however, the increase in pH had no effect on zinc (galva-
nized iron) pipe. In a review of the first round of LCR monitoring data, it was observed that
no copper action level (AL) exceedances were reported for systems having a pH greater
than approximately 8.0 (Schock and Lytle 1995).
In a survey of nearly 400 US drinking water utilities, Dodrill and Edwards (1995)
observed that for utilities not adding a phosphate inhibitor at pH less than 8.4, lead release
was significantly lower when alkalinity was 30–74 mg/L compared to alkalinity less than
30 mg/L. With regard to copper release, it was reported that problems meeting the copper
AL occurred under two water quality scenarios: (1) pH less than 7.0 and alkalinity less
than 30 mg/L and (2) pH less than 7.8 and alkalinity greater than 90 mg/L. In a review of
the results of the first round of LCR monitoring data, Schock and Lytle (1995) noted that
copper AL exceedances tended to be highest at low alkalinities (<25 mg/L as CaCO3) and
increasingly greater over 75 mg/L as CaCO3.
A direct correlation between red water incidents and finished water alkalinity has
also been observed. Horsley et al. (1998) noted that the frequency of red water occur-
rence increased significantly when the finished water alkalinity dropped below 60 mg/L.
In this case, the authors concluded that the optimum finished water alkalinity is likely to
vary with finished water chemistry and system configuration (for example, the types of
pipe present and other potential corrosion concerns, such as lead). However, alkalinity
adjustment should be considered when attempting to address red water problems. Reiber
(2006) observed that iron release more than doubled following a decrease in alkalinity
from approximately 30 mg/L to 10 mg/L in a pilot distribution system at a pH of approxi-
mately 7.5. Imran et al. (2005) determined decreases in alkalinity concentrations resulted
in increases in red water occurrence in pilot distribution systems in which calcite forma-
tion was used as a primary means of corrosion control.
Portland Water District, Maine (PWD), experienced firsthand the significance of DIC
and the impact on buffer intensity with regard to lead corrosion control (Boissonneault
1994). In the early 1990s, like much of the United States, PWD was struggling to comply
with the LCR. At the time, PWD utilized a low alkalinity (<5 mg/L), low DIC (<2 mg C/L)
source water that had a poor buffering capacity and was extremely corrosive. Corro-
sion control at the time included pH adjustment to approximately 7.6 and addition of a
zinc orthophosphate for the control of red water. Multiple corrosion control strategies,
including pH adjustment, pH adjustment plus zinc orthophosphate, and pH adjustment
plus zinc polyphosphate were tried. None of these strategies were effective for the control
of lead corrosion. It was ultimately determined that the preferred lead control strategy
was increasing the DIC using carbon dioxide or sodium bicarbonate to increase the buf-
fer intensity and to aid in maintaining a stable finished water pH, combined with zinc
orthophosphate addition and a target finished water pH of 7.6.
It has been suggested that ion exchange softening is likely to create a more corrosive
water because of the removal of calcium and subsequent increase in TDS that results from
its use. However, it has been observed that ion exchange softening has little effect on lead,
M58 book.indb 37 11/17/2010 4:24:47 PM

copper, and galvanized iron corrosion rates (Sorg et al. 1999) because the ion exchange
process does not significantly alter the parameters that have the most significant impact
on corrosion: pH, alkalinity, and, most importantly, DIC. This observation demonstrates
the importance of DIC in implementing and maintaining an effective corrosion control
strategy.
Oxidants
The presence of oxidants, such as dissolved oxygen, chlorine, and chloramine, can signifi-
cantly impact internal corrosion. Many instances have been reported where failure to con-
sider the presence or absence of a particular oxidant and its potential impact on corrosion
has resulted in LCR action level exceedances or increases in red water complaints.
Dissolved oxygen. The presence of any dissolved gas (e.g., CO2) has the potential
to impact corrosion. However, of the several potential dissolved gases found in drinking
water, dissolved oxygen is most likely to impact corrosion. Oxygen is only slightly soluble
in water, seldom reaching concentrations exceeding 15 mg/L. Increases in dissolved oxy-
gen can affect the solubility of iron, manganese, lead, and copper.
Dissolved oxygen may react with ferrous (Fe2+) ions in water and convert them to
ferric (Fe3+) ions, which can form ferric hydroxide and result in red water complaints
(Schock 1999). Imran et al. (2005) identified dissolved oxygen as a contributing factor to
red water episodes in a pilot distribution system under various water quality conditions.
In these studies, increases in dissolved oxygen concentrations resulted in increases in red
water occurrence.
Dissolved oxygen may also increase pipe tuberculation and pitting corrosion. Schock
et al. (1998) noted that copper pipe in water containing oxidizing agents will continue to
corrode either until all of the oxygen is depleted or until precipitated oxide films arrest the
rate of corrosion. However, substantial dissolved oxygen concentrations may also result
in the formation of more protective metal oxides and reduce corrosion.
Oxidant type and residual. Free chlorine residuals can impact the metal oxida-
tion rates and the nature of scales that form on the interior of distribution system piping.
Chlorine addition to soft waters has been observed to both increase and decrease corro-
sion rates for copper (Boulay and Edwards 2001). Several researchers have also suggested
(Edwards 2004; Lytle and Schock 2005) that elemental chlorine promotes formation of
lead oxides which contribute to the formation of passivated layers inside distribution sys-
tem piping. That is, chlorine is such a strong oxidant that lead loses four electrons instead
of two. Then, lead can combine with oxygen in the water to form highly insoluble com-
pounds that protect the pipe from the water.
This phenomenon between chlorine and lead may also be seen in the data from one
pipe loop study (Cantor et al. 2003). In that study, the addition of chlorine at a dose of
0.2 mg/L actually reduced lead corrosion but resulted in increased corrosion of copper
and iron.
Pisigan and Singley (1987) observed that corrosion rates of both mild steel and
copper were higher in the presence of free chlorine than in its absence. It was further
observed that the corrosion rates of aerated water were less than those of deaerated water
containing free chlorine, suggesting that free chlorine is more corrosive than dissolved
oxygen. Aerated water containing free chlorine was the most corrosive of the conditions
evaluated. Stone et al. (1987) observed free chlorine residual concentration had a signifi-
cant impact on both aged and fresh (new) copper pipe corrosion but did not impact zinc
(galvanized iron) corrosion.
Oxidation–reduction potential. ORP, also frequently referred to as redox poten-
tial, is a measure of water’s capability to oxidize and is reported as electrical potential
(volts, V, or millivolts, mV). Disinfectants are, by their very nature, oxidants and as such
adding a disinfectant will increase ORP in the water.
M58 book.indb 38 11/17/2010 4:24:47 PM

Some disinfectants are stronger oxidants than others. For example, chlorine can
quickly oxidize ferrous (dissolved) iron. Chloramine, which is becoming a more common
disinfectant used to reduce DBPs, will oxidize ferrous iron but not as readily as chlorine.
Therefore, utilities that switch from chlorine to chloramine will reduce ORP values in the
distribution system.
The Washington Aqueduct utility (that is, DC Water) that serves Washington, D.C.,
switched from using free chlorine for secondary disinfection to chloramine in November
2000, which appeared to have had some role in the increased lead concentrations found
within the system. Bench-scale experiments (Edwards and Dudi 2004) seemed to confirm
that a switch from free chlorine to chloramine can accelerate the dissolution of lead under
conditions similar to those in Washington D.C., particularly where brass and bronze in-line
devices (faucets, fixtures, water meters, and so on) are present. However, although some
have implicated chloramine as the cause of the elevated lead levels in Washington D.C.,
experts thus far have suspected that free chlorine actually served as a form of corrosion
control treatment (Reiber 2004). The hypothesis is that it was not chloramine addition per
se that was contributing to lead leaching but rather the absence of the free chlorine previ-
ously employed. Other researchers have shown that chloramine also altered the oxidation
state of existing lead scales, which resulted in a shift in mineral species from Pb4+ to Pb2+
(Schock and Giani 2005; Renner 2006; Reiber 2004). Prior to chloramine conversion, lead
scales in the Washington, D.C., system had primarily consisted of less soluble Pb4+ species.
However, the change in redox potential caused an equilibrium shift to more soluble Pb2+,
which resulted in an increase in tap lead concentrations. Note that higher ORPs have been
found to be more effective for corrosion control because higher ORPs generally result in
the formation of a more stable metal complex (USEPA 2007). A much more detailed discus-
sion of the experience in Washington, D.C., is presented in Appendix B.
Typical drinking water ORP ranges from about 400 mV to 600 mV. However, chlo-
rine concentrations above 2.5 mg/L can raise ORP levels above 600 mV. Higher ORP levels
will increase oxidation rates in the drinking water. Schock and Giani (2005) were able to
demonstrate that scale formation on pipe walls may change with changes in ORP. Utilities
that add a stronger oxidizing disinfectant such as chlorine and switch to a lower oxidizing
disinfectant such as chloramine would expect to see a decrease in ORP. Large decreases
in ORP (> 100 mV) may see existing scales begin to weaken, dissolve, or change chemical
form (e.g., Pb4+ to Pb2+). This behavior includes lead oxide scales that might have formed
on existing lead pipes due to high ORP values.
Pourbaix diagrams show the impact of electrochemical potential (or ORP) and pH
on metal speciation and mineral formation and are frequently referred to as potential-pH
diagrams or Eh-pH diagrams. Figure 3-4 presents a typical Pourbaix diagram for an iron-
water system. Similar diagrams are available for lead and copper and for a variety of water
environments. Pourbaix diagrams do not predict corrosivity or metal solubility but can
provide insight into metals speciation and mineral stability. Using Figure 3-4, for example,
if the potential of a chlorinated water is 1.5 volts and the pH is 8.0, it can be predicted that
the iron species present will be predominantly Fe(OH)3 precipitate, which is likely to be
the most stable precipitate from the perspective of maintaining an iron scale on distribu-
tion pipe walls.
Anions
The presence of some anions (e.g., Cl–, S –, SO4 –, NO3 –) can play a significant role in the
internal corrosion of distribution system and home plumbing materials. This section
briefly discusses some of the more significant anions with regard to corrosion and their
potential impacts on corrosion.
M58 book.indb 39 11/17/2010 4:24:47 PM

Source: Schock et al. 1999.
Figure 3-4 Iron Pourbaix diagram for a carbonate-containing water at 25ºC and I = 0. Stability
fields are shown for dissolved iron species activities of 0.1 mg/L(—) and 1.0 mg/L (– – –). Dissolved
carbonate species concentrations are 4.8 mg C/L (4 ×10-4 M)
Sulfide. The presence of sulfide has been demonstrated to significantly increase cor-
rosion rates and the occurrence of pinhole leaks in copper plumbing. Jacobs et al. (1998)
demonstrated that sulfide-containing waters were substantially more aggressive with
regard to copper. Even following elimination of the sulfide from the source water and
removal of previous pipe scales, pipes previously exposed to sulfide corroded at substan-
tially higher rates than pipes never exposed to sulfide. Sulfide is also a by-product of reduc-
tion of sulfate by SRB and can be an indicator of MIC (Jacobs et al. 1998).
Chloride. Chloride has been demonstrated to be a critical anion in the initiation
of copper corrosion and pitting (Ives and Rawson 1962). However, Edwards et al. (1994)
found that copper surfaces passivated in the presence of chloride at pH ≥ 7.0 in aged piping
systems.
Imran et al. (2005) determined increases in chloride concentrations in the range of
22 to 70 mg/L resulted in increases in red water occurrence in pilot distribution systems.
Lytle et al. (2003) observed that total iron release increased by nearly 50 percent in the
presence of 100 mg/L of chloride. Addition of orthophosphate at a concentration of 3 mg/L
as PO4 was able to reduce total iron release to normal levels. Following discontinuation of
orthophosphate addition, total iron release again increased but not significantly, suggest-
ing some iron-phosphate mechanism that produced lasting benefits (Lytle et al. 2003).
Sulfate. Imran et al. (2005) determined increases in sulfate concentrations resulted
in increases in red water occurrence in pilot distribution systems. Edwards et al. (1994)
demonstrated that sulfate was the most significant of the anions studied (including car-
bonate, chloride, perchlorate, and nitrate) with regard to copper corrosion rates. In these
studies, water with high sulfate concentrations (100 mg/L) exhibited increased corrosion
rates with aging.
M58 book.indb 40 11/17/2010 4:24:47 PM

Chloride-to-sulfate ratio. The chloride-to-sulfate mass ratio (Cl–:SO42–) has been

observed to have a significant impact on the leaching of lead from soldered joints and brass
(Triantafyllidou and Edwards 2006) and in lead plumbing (Edwards et al. 1999). Changes in
coagulant to improve NOM removal and to ultimately reduce DBPs may also cause shifts in
lead and copper solubility, depending on the type of corrosion control treatment employed.
Edwards et al. (1999) reported that in a utility survey, 100 percent of the utilities with
Cl–:SO42– ratios less than 0.58 met the 0.015 mg/L Pb action level. However, of those facili-
ties with Cl–:SO42– ratios greater than 0.58, only 36 percent met the action level.
Carlson et al. (2000) noted that one utility switching from alum to ferric chloride
to enhance total organic carbon (TOC) removal and reduce distribution system tri-
halomethane (THM) concentrations resulted in that utility exceeding the lead AL. The
change in coagulants resulted in an increase in the Cl–:SO42– ratio; however, all other key
corrosion parameters remained relatively constant. Table 3-2 summarizes the impact of
coagulant changes on system Pb levels for this utility.
The city of Columbus (Ohio) Division of Water (DOW) experienced similar detrimen-
tal effects on lead corrosion following a switch from alum to ferric chloride. Figure 3-5
shows the Cl–:SO42– ratio and 90th percentile lead concentration for the DOW service area.
A “system average” Cl–:SO42– ratio is presented because DOW operates three plants—two
surface water and one groundwater. The surface water plants were switched over from
alum to ferric chloride in July 1995 and September 1996. The groundwater plant is purely
a lime softening plant and does not use a coagulant.
The issues faced by Stafford County (Virginia) Department of Utilities are similar
to those of many utilities across the country and exemplify the difficulties of complying
with multiple drinking water regulations and, in this case, wastewater treatment require-
ments. The county operates two conventional surface water treatment plants. Prior to
2001, both of the plants used alum for coagulation and zinc orthophosphate was used for
corrosion control. In 2001, concerns about zinc concentrations at the county wastewater
treatment plants motivated a change to an orthophosphate–polyphosphate blend for cor-
rosion control. In 2003, when the Stage 1 Disinfectants and Disinfection By-Products Rule
(D/DBPR) required additional TOC removal at one of the plants, the decision was made
to switch from alum to ferric chloride at that treatment plant. After the switches to the
orthophosphate–polyphosphate blend and ferric chloride, 4 of 32 tap samples taken in
2003 had lead concentrations exceeding the 15 µg/L action level. In 2004, 12 of 60 samples
had lead concentrations exceeding the action level. The county has since changed from
the orthophosphate–polyphosphate blend to straight orthophosphate and is in the process
of converting from ferric chloride to ferric sulfate.
While it is unclear which treatment change had the most significant impact on the
increased lead corrosion (the change to polyphosphate or switch to ferric chloride),
both factors have previously been identified as responsible for increases in lead corro-
sion (Carlson et al. 2000; Edwards and McNeill 2002). Edwards and McNeill (2002) noted
that polyphosphate can be effective for the control of particulate iron and manganese in
finished water but is generally not recommended for lead and copper corrosion control.
Another of the factors that appears to be critical in lead corrosion is the Cl–:SO42– ratio.
Carlson et al. (2000) and the city of Columbus, Ohio, noted increases in lead concentra-
tions following changes from alum to ferric chloride. After conversion to ferric chlo-
ride, Stafford County observed an increase in the Cl–:SO42– ratio from 0.29 to 0.38 prior
to the switch to 4.75 after the switch. The change from ferric chloride to ferric sulfate
reduced the Cl–:SO42– ratio to well below the levels present prior to 2003 and reduced
90th percentile lead concentrations within 1 to 2 months (Edwards and McNeill 2002).
M58 book.indb 41 11/17/2010 4:24:47 PM

Table 3-2 Impact of coagulant changes on lead corrosion

Primary Coagulant
Parameter Alum Ferric Chloride
pH* 7.5 7.4
Alkalinity, mg/L as CaCO3* 75.0 70.0
Sulfate, mg/L* 74.0 34.0
Chloride, mg/L* 24.0 38.0
Chloride sulfate ratio 0.32 1.11
TOC, mg/L 2.5 1.9
Total THMs, µg/L 65.0 40.0
Number of distribution samples > than 15 µg/L Pb 0/71 12/71
*Finished water.
Source: Carlson et al. 2000.
1.00 35
0.90 System Avg Cl:SO4

90th Percentile Pb 30
0.80
25
0.70
90th Percentile Pb
0.60 Cl:SO4 = 0.58
Cl:SO4 Ratio
20
0.50
15
0.40 Sept 97: Stopped
FeCl3 Addition
0.30 July 95:Began
10
FeCl3 Addition
0.20
5
0.10
0.00 0
Jan-94
Apr-94
Jul-94
Jan-95
Jul-96
Jan-97
Oct-94
Apr-95
Jul-95
Jan-96
Oct-95
Apr-96
Oct-96
Apr-97
Jul-97
Oct-97
Date
Figure 3-5 Impact of chloride-to-sulfate ratio on lead corrosion
Other anions. Edwards et al. (1994) found that copper surfaces activated
(i.e., increased corrosion rates with aging) in the presence of nitrate, perchlorate, and sul-
fate over the pH range 5.5 to 10. It was suggested that the relative aggressiveness of various
anions toward copper after aging was
HCO3 – > SO42– > NO3 – > ClO4 – > Cl– (at pH 7.0)
SO42– ≥ ClO4 – > NO3 – > HCO3 – > Cl– (at pH 8.5)
M58 book.indb 42 11/17/2010 4:24:48 PM

Corrosion Inhibitors
Currently, the use of corrosion control chemicals is common in the drinking water industry.
However, each chemical has different properties and may not truly inhibit corrosion. Some
simply mask the effects of corrosion to prevent aesthetic problems, such as red or black
water. The most common inhibitors used in the drinking water industry are orthophos-
phates, polyphosphates, blended phosphates, and silicates. The phosphate-based inhibi-
tors are by far the most commonly used chemical inhibitors. Chapter 4 discusses the use
of corrosion inhibitors in greater detail.
Orthophosphate. The predominant phosphate species in corrosion inhibitor chemi-
cals has a significant impact on its performance as a corrosion inhibitor. Orthophosphate
(PO4 –3) added as a corrosion control treatment chemical can combine with lead and copper
in plumbing materials to form a wide variety of insoluble compounds, which is an effective
corrosion control strategy known as passivation. Chapter 2 discusses the mechanism of
how inhibitors reduce corrosion. In laboratory studies, Edwards and McNeill (2002) found
that orthophosphate dosing at 1 mg P/L was able to reduce soluble lead release by approxi-
mately 70 percent in most cases. However, it was also observed in that, in the case of some
new pipes, lead release increased. The authors theorized this finding may have been due
to the fact that pH increase resulting from orthophosphate stagnation was less in new
pipes. Reiber (2006) noted that orthophosphate was effective for hardening existing iron
scales at pH 7.4–7.8 and can help to prevent red water occurrences. Similar results were
reported by Lytle et al. (2003), who observed that total iron released remained low follow-
ing discontinuation of orthophosphate addition due to the formation of iron-phosphorous
solids in the scales, thereby reducing the solubility of ferrous iron and/or decreasing the
permeability of the scales.
Polyphosphate. Reiber (2006) noted that polyphosphate can be effective for the con-
trol of red water by mobilizing friable iron and helping to stabilize the corrosion scale core.
In one utility, the use of polyphosphates was credited with reducing the number of red and
black water complaints from as many as 35 per month to less than 5 per month in less than
1 month (Arweiler et al. 2003). Holm and Schock (1991) also noted that the complexation
properties of polyphosphates can be effective for the control of iron (red water), manga-
nese (black water), and carbonate scaling but may also significantly increase the solubility
of lead in home plumbing systems. Cantor (2000) found that polyphosphate increased both
tap lead and copper concentrations in high DIC waters, while orthophosphate was gener-
ally effective for the control of lead and copper. Cantor also observed that the impacts of
polyphosphate on copper corrosion may be more significant in newer homes.
Dodrill and Edwards (1995) reported that the 90th percentile lead release when dos-
ing polyphosphate at 1 mg P/L was more than double that when dosing orthophosphate in
utilities with alkalinities less than 30 mg/L and pH less than 7.4. Further, under similar pH
and alkalinity conditions, lead release using polyphosphates was greater than when dos-
ing no inhibitor at all. In the same alkalinity range but at pH 7.4 to 7.8, polyphosphate actu-
ally performed better than orthophosphate, and both performed better than having added
no inhibitor at all. At alkalinities 30 to 74 mg/L and pH 7.4 to 7.8, there did not appear to be
any significant difference in orthophosphate and polyphosphate performance.
Polyphosphates will revert to orthophosphate over time in the distribution system.
The half-lives of most simple polyphosphates are on the order of a few days (Clesceri and
Lee 1965). In pipe rig experiments, Holm and Edwards (2003) confirmed these results and
found that 20 to 95 percent reversion occurred in as few as 3 days. For this reason, the
results of many polyphosphate studies may be confounded by the presence of orthophos-
phate (McNeill and Edwards 2001). Determination of the fractions of orthophosphate
and polyphosphate present can be assessed by measuring the total phosphorous and
orthophosphate concentrations. The relationship between the three parameters is
M58 book.indb 43 11/17/2010 4:24:48 PM

PP = PT − PO
Where:
PP = polyphosphate fraction
PT = total phosphate concentration
PO = orthophosphate fraction
Zinc. In the 1960s, inhibitor manufacturers began blending orthophosphates and

polyphosphates with zinc. It was claimed that the zinc accelerated polyphosphate film for-
mation or that zinc orthophosphate or zinc polyphosphate films were superior for inhibit-
ing corrosion (McNeill and Edwards 2001). Several recent studies have found no additional
benefit as a result of the presence of zinc (McNeill and Edwards 2000; Williams 1990).
Sodium hexametaphosphate. Sodium hexametaphosphate (SHMP), also known as
glassy phosphate or metaphosphate glass, is a sequestering agent and has been shown to
be effective for red water control (McNeill and Edwards 2001), but SHMP is not typically
effective as a lead corrosion inhibitor. Under controlled laboratory conditions, Edwards
et al. (2002) observed that soluble lead concentration increased by approximately 1.6 mg/L
for each 1 mg/L increase in SHMP.
Silicates. Sodium silicate (Na2SiO3) is used primarily as a sequestering agent for
the control of iron and manganese. The primary corrosion protection offered by silicates
appears to be the resulting increase in pH and formation of a protective film (Vik et al. 1996).
For this reason, silicates are most likely to be used in soft waters of low pH and high oxygen
concentrations, although they have been successfully applied in varying water qualities.
The effectiveness of silicate inhibitors may be limited by the presence of multivalent
cations such as calcium and magnesium (Robinson et al. 1987). With respect to lead and
copper, sodium silicates are basic and may result in an increase in pH, which is usually
beneficial from a lead and copper control perspective. It has also been suggested that
silicates form a thin protective layer that acts as a diffusion barrier to lead and copper
corrosion (Stericker 1945). The dose required for passivation of lead and copper by sodium
silicate solution is typically much higher (20–30 mg/L as sodium silicate) than that of the
phosphate-based inhibitors (1–2 mg/L as P). In low pH waters, it may be necessary to use
high silicate doses, and it may be more economical to raise the pH using caustic soda or
soda ash prior to silicate addition (Vik et al. 1996).
Significant reductions in 90th percentile lead and copper concentrations have been
reported following the addition of silicates (Schock et al. 1998; Schock et al. 2005). In
one study, 90th percentile lead concentrations were reduced from 0.077 mg/L to 0.002
mg/L, and 90th percentile copper concentrations were reduced from 5.87 mg/L to 0.27
mg/L following a change from a polyphosphate inhibitor to sodium silicate. The silicate
dose ranged from 25 to 30 mg/L and resulted in an increase in the treated water pH from
approximately 6.1 to 7.5.
Biostability
The biostability of a water, or its propensity to support microbiological activity, can affect
corrosion in a number of ways. Microorganisms, primarily bacteria, are present in nearly
every distribution system. It has been theorized that system biofilms may serve as a pro-
tective barrier to corrosion (Abernathy and Camper 1998). However, biofilms may also
produce a differential potential, leading to localized changes in oxygen concentration and
electrical potential (Lee et al. 1980). Bacteria may also consume oxygen. Alkalinity lead-
ing to localized pH gradients, which may affect corrosion, can produce corrosive metabo-
lites such as hydrogen sulfide or iron phosphide (Tuovinen et al. 1980).
M58 book.indb 44 11/17/2010 4:24:48 PM

Biocorrosion, or MIC, of copper pipe may be the result of (1) the production of acid
metabolites on the metal surface, (2) binding of copper by bacterial metabolites (extra-
cellular polysaccharides) that inhibits the natural bactericidal effects of copper, and (3)
alteration of the normal cupric oxide film from the incorporation of bacterial cells and
polymers (Bremer et al. 2001).
A number of factors can contribute to MIC of copper pipe. These factors include
(Bremer et al. 2001)
• Use of soft water with low pH (neutral or neutral-alkaline)
• High suspended solids and AOC content
• Long-term or periodic stagnation in the pipe, which produces widely fluctuating
oxygen concentrations
• Dead-ends or long horizontal runs that are susceptible to sediment accumulation
• Low or nonexistent levels of chlorine
• Maintenance of water temperatures that promote rapid growth and activity of nat-
urally occurring bacteria that form biofilms on the pipe wall
• Lack of an adequate monitoring program to periodically evaluate water quality
and pipe wall conditions
Assimilable organic carbon. AOC is the fraction of biodegradable organic carbon
consisting of small organic compounds that can be used easily by microorganisms as a
food source. Generally, AOC makes up only 0.1 to 9.0 percent of the TOC present in a water
(Escobar and Randall 2001). Van der Kooij (1992) reported that the AOC level must be less
than 10 µg/L to prevent microbiological activity in the absence of a disinfectant.
Ozonation may break down larger organic molecules into smaller molecules result-
ing in an increase in AOC concentrations. Failure to follow ozonation with biologically
active filters can result in increases in distribution system AOC concentrations. It has been
theorized that in at least one instance this increase in AOC resulted in an increase in the
occurrence of MIC and pinhole leaks in copper pipes (Cantor et al. 2003). Subsequent eval-
uation in the same system determined that increased disinfectant residuals eliminated
copper pipe failures and appeared to have confirmed the occurrence of MIC (Cantor et al.
2006).
Disinfectant residual. Reiber et al. (1987) observed that corrosion rates varied sig-
nificantly with fluctuations in distribution system free-chlorine residual concentrations in
both aged and new copper pipes in a low alkalinity (15–17 mg/L as CaCO3) water. Bremer
et al. (2001) noted that MIC of copper was predominant in pipes with low or no disinfectant
present. Zhang et al. (2002) observed that maintaining a total chlorine (i.e., chloramine)
residual greater than 2 mg/L was essential for maintaining biostability and controlling
microbial regrowth. It was also noted that even in waters with moderate to high AOC con-
centrations, biostability could be maintained given proper maintenance of a disinfectant
residual. Such factors are critical to the prevention of MIC of copper pipe.
Other bacterial nutrients. Ammonia, sulfate, nitrate, and phosphorus can serve
as potential nutrient sources for a variety of bacteria commonly found in distribution and
home plumbing systems.
Ammonia is a nutrient source for ammonia-oxidizing bacteria (AOB) and is the
precursor to nitrification. AOB oxidizes ammonia to form nitrite (partial nitrification),
which may in turn serve as a nutrient source for nitrite-oxidizing bacteria (NOB). NOB
oxidize nitrite to form nitrate (complete nitrification). The nitrification process itself con-
sumes alkalinity and may result in localized pH depression, increasing the potential for
corrosion.
M58 book.indb 45 11/17/2010 4:24:48 PM

Sulfate is a nutrient source for SRB and has been demonstrated to be a key factor in
MIC of copper piping. Further, in oxygen deficient conditions SRB reduce sulfate to sul-
fide, which is a critical anion in the pitting corrosion of copper pipe (Jacobs et al. 1998).
Similarly, phosphorous is a potential nutrient source and may increase microbiologi-
cal activity in distribution and home plumbing systems. Increased microbiological activity
increases the potential for MIC.
Heterotrophic plate count. Heterotrophic bacteria utilize organic carbon present
in distribution systems and household plumbing as a source of energy. An HPC is an enu-
meration method for heterotrophic bacteria and is used as an indicator of microbiological
water quality in the drinking water industry. There are two common HPC methods, the
standard plate count agar and R2A agar method. It is recommended that the R2A agar
method be used for all HPC analyses because it is more sensitive and is likely to result in
more accurate depiction of microbiological activity (Smith 2006).
Elevated HPCs compared to baseline concentrations indicate increased microbiolog-
ical activity is likely to occur in areas of the distribution system or home plumbing system
in which there have been significant losses in disinfectant residual, where AOC levels are
sufficient to encourage microbial regrowth, or in which nitrification is occurring. Can-
tor et al. (2003) observed MIC was occurring in areas of a distribution system with AOC
concentrations between approximately 1.4 and 126 µg/L in which the total chlorine (i.e.,
monochloramine) residual concentration was negligible. In addition, areas of chloram-
inated distribution systems where no residual is present are known to be more susceptible
to nitrification (Fleming et al. 2005).
NOM
NOM has been found to reduce the corrosion rates of both galvanized steel and cast-iron
pipe (Sontheimer et al. 1981; Larson 1966). NOM has also been found to produce a more
protective scale in iron pipes by reducing ferric colloids to soluble ferrous iron (Campbell
and Turner 1983). One reason is that the presence of humic and fulvic acids promotes the
formation of protective ferrous iron scales (Benjamin et al. 1996).
Schock et al. (1996) reported varied impacts of NOM on lead and copper solubility. In
some instances, NOM was observed to form soluble organic complexes with lead, result-
ing in an increase in dissolved lead concentrations. Burlingame et al. (2006) noted that
the formation of Pb(II)–NOM complexes increases the solubility of lead such that as NOM
increases so does the occurrence of soluble lead. Similarly, Korshin et al. (2005) demon-
strated that the presence of NOM can prevent or hinder the formation of lead carbonate
species, resulting in increased lead concentrations. In other cases, NOM adsorbed or other-
wise adhered to the inside of lead pipes, decreasing lead solubility.
Similar results have been reported for copper. Ferguson et al. (1996) reported that
copper can form complexes with organic material resulting in lower dissolved copper con-
centrations. Edwards et al. (1994) noted that copper corrosion rates were substantially
lower in the presence of 2 mg/L NOM than in its absence. Other researchers have noted
that NOM increases copper corrosion, particularly following one night of stagnation (Broo
et al. 1999). In such a case, it is possible that MIC was occurring; however, this study did
not investigate MIC as a possible cause of increased copper release.
Metals
The presence of trace metals, such as magnesium and zinc, may inhibit calcite formation
on pipe surfaces and favor the deposition of more soluble forms of CaCO3, which are less
likely to prevent interactions between the pipe and the water (LeChevallier et al. 1993).
The presence of copper, as a result of dissolution of plumbing materials or as a result of
copper sulfate addition as an algaecide, may cause rapid corrosion of galvanized pipe.
M58 book.indb 46 11/17/2010 4:24:48 PM

Other trace metals, such as aluminum and magnesium, may also impact corrosion and
influence passivating layer and scale formation.
Total metals and dissolved metals. When conducting metals analyses, it is impor-
tant to distinguish between total metals and dissolved metals, for example, total lead and
dissolved lead. The reasons associated with the occurrence of dissolved metals, as opposed
to particulate metals, may be vastly different. Similarly, it is possible to see changes in the
fraction of dissolved and particulate metals but see no change in total metals concentra-
tion. Cantor (2006) noted the importance of distinguishing between dissolved and particu-
late metals when trying to determine the potential source of corrosion-related problems.
Boyd et al. (2004) found that colloidal lead concentrations represented 38 to 54 percent of
total lead and dissolved lead concentrations were 41 to 60 percent of total lead concentra-
tions after stabilization in simulated partial lead service line replacement studies.
Madison Water Utility exceeded the action level for lead in 1992 with a 90th percen-
tile value of 16 µg/L. After it was determined that the source of the lead was lead service
lines and various treatment and other options were evaluated, it was decided to imple-
ment an aggressive lead service line replacement program. To evaluate the effectiveness
of the program, first-draw samples were collected from 60 homes at which lead service
line replacement had been completed. Surprisingly, the 90th percentile value of these
samples for total lead was 22 µg/L, or an increase of 6 µg/L over the pre-LSL replacement
value. However, the 90th percentile value of the samples for dissolved lead was just 5 µg/L,
leading Madison to conclude that particulate lead was the source of elevated lead levels
(Cantor 2006). This example illustrates the importance in differentiating between total
metals and dissolved metals concentrations. After determining the problem to be random
particulate lead released both before and after lead service line replacement, Madison
warned consumers to continue allowing some water to run to waste before drinking for
several years after lead service line replacement.
Aluminum. Aluminum may be a widespread component of distribution films and act
as a barrier to corrosion and metals release. On the one hand, Lauer and Lohman (1994)
found aluminum films to significantly reduce lead leaching. On the other hand, Scardina
(2006) noted significant increases in the occurrence of copper pitting in plumbing sys-
tems that coincided with increased dissolved aluminum concentrations in the distribu-
tion system. Aluminum has also been found in conglomerated precipitates that appear to
form around copper pinhole leaks (Rushing and Edwards 2002). However, a direct correla-
tion between aluminum concentrations and the frequency of pinhole leaks has not been
determined.
This parameter is also important in those systems that use aluminum-based coagu-
lants, such as alum. Improper coagulation process may result in elevated aluminum levels
entering into the distribution system. Elevated aluminum levels become important when
adding a phosphate-based inhibitor for corrosion treatment. Aluminum will readily bind to
orthophosphate to form an aluminum phosphate precipitate (AlPO4). Too much aluminum
(>0.1 mg/L) may impede the ability of phosphate to react with lead and copper. In addition,
aluminum phosphate can cause a precipitate that may increase “cloudy water” complaints
at the customer tap (AWWA 2005). This precipitate is very small and has a low settleabil-
ity rate, which can allow it to settle in low flow areas at the end of distribution systems.
Iron and manganese. In addition to the potential of iron and manganese to con-
tribute to colored water episodes, iron and manganese pose other concerns related to
corrosion. Iron is an important parameter for several reasons. Some bacteria utilize iron
as a food source, which can result in increased microbial activity and increased potential
for MIC to occur. In addition, both iron and manganese precipitates on pipe surfaces may
form a physical barrier between the microorganisms and the drinking water disinfectant
allowing microbiological activity to increase and thus cause an increased potential for
localized corrosion (Figure 3-6).
M58 book.indb 47 11/17/2010 4:24:48 PM

Figure 3-6 Precipitated iron as a protective barrier for microorganisms
Recent studies have also shown that under certain conditions, manganese may be a
significant component of lead pipe scale. In one utility, it was observed that manganese
scales had adsorbed lead and other metals. In this system, even after lead service line
replacement, manganese scales were released, resulting in increased lead and other met-
als concentrations at the consumers’ taps. It is speculated that a reduction in ORP values
may cause the manganese–lead scale to weaken or become soluble, increasing particulate
lead concentrations (Schock et al. 2006; Maynard and Mast 2006).
Turbidity and Color

Turbidity is the measurement of the amount of light diffraction in a sample caused by par-
ticulates and sediments; therefore, the higher the turbidity, the higher the level of particu-
lates in the water. These particulates may cause water discoloration or eventually settle
on pipe walls, providing increased opportunity for microbial regrowth in the distribution
system. Apparent color compares the color of water to that of deionized water.
Turbidity and color are relevant to corrosion for several reasons. Both will increase
during red water or other aesthetic events. In addition, these parameters, combined with
total metals and dissolved metals analyses, may help to better characterize a corrosion-
related event. For example, an increase in the number of red water complaints accompa-
nied by an increase in turbidity and particulate iron but with no significant change in total
iron concentrations may indicate that something is occurring in the distribution system
to oxidize ferrous iron already present. Furthermore, increases in turbidity as a result
of iron particles may also result in increases in particulate lead concentrations due to
adsorption of lead onto the iron scales (Kirchner 2009). This data, when combined with
chlorine residual, ORP, and total iron and phosphate (total phosphate, orthophosphate,
and/or polyphosphate), may help to understand why the red water event occurred.
Other Considerations
In addition to water quality, a number of other factors not only can influence corrosion and
increased metals release but also can be indicative of corrosion-related problems. Hydrau-
lic factors, customer complaints, and unaccounted-for water should also be considered
and may be useful in identifying the location and cause of water quality problems related
to corrosion.
Hydraulic factors. Several researchers have observed that corrosion rates of lead,
copper, and iron are likely to be highest in areas of the distribution system with the longest
hydraulic residence time, or water age (Edwards et al. 1994; Imran et al. 2005). Several rea-
sons are possible, the most significant being that disinfectant residuals tend to be lower
M58 book.indb 48 11/17/2010 4:24:50 PM

in areas of the distribution system with higher water age. Theoretically, low disinfectant
residuals can reduce the ORP of the water and can cause shifts in metal species in exist-
ing pipe scales, resulting in an increase in dissolved metal concentrations. Further, low
or nonexistent disinfectant residuals increase the potential for MIC. It has been observed
that low flow or stagnant areas of the distribution system or home plumbing systems are
more prone to MIC (Cantor et al. 2003; Bremer et al. 2001).
It has also been observed that stable iron corrosion deposits may be disrupted by
changes in flow (Smith et al. 1998). Pisigan and Singley (1987) found that mild steel and
copper corrosion increased significantly as flow rates increased. It was observed that with
time, passivating layers on the pipes could form; however, further increases in the flow rates
or turbulent flow conditions (as are typically found in distribution systems) may disrupt the
protective surface. Corrosion rates of copper piping are not anticipated to be excessive if
the flow rates are less than 4 to 5 ft/s (Obrecht and Quill 1960; Obrecht and Quill 1961).
Water quality customer complaints. Changes in aesthetic parameters, as indi-
cated by monitoring or increases in customer complaints, may indicate conditions are
such that corrosion control treatment effectiveness has been compromised. For example,
customers may notice a change in taste or odor in old or stagnant water. It has been dem-
onstrated that some chemicals added to prevent iron release and red water problems (e.g.,
polyphosphates) are not effective for control of lead and copper corrosion, particularly in
newer homes.
Unaccounted-for water. Unaccounted-for water, or water losses, may also be an
indicator of widespread corrosion control failure. Significant increases in unaccounted-
for water may result from failure of both distribution system piping and home plumbing
systems. In Brown Deer, Wis., an increase in unaccounted-for water from 3.4 percent to 9.4
percent over a four-year period coincided with a substantial increase in the occurrence of
copper pipe failures in home plumbing systems (Cantor et al. 2003). Subsequent evaluation
of this system, after remedying the copper pipe corrosion problem, showed unaccounted-
for water had dropped to 2 percent (Cantor et al. 2006).
DEVELOPING A WATER QUALITY MONITORING PROGRAM_______

Importance of Establishing Baseline Water Quality
It is necessary to establish baseline water quality not only to identify the conditions under
which corrosion control is most effective but also to determine the cause of corrosion-
related episodes, such as red water events and increased tap lead concentrations. Baseline
water quality data should be established for the source water, treatment plant, entry point,
distribution system, and household plumbing.
Many systems may have historical data that can help provide initial baseline data
when a monitoring program is initiated. Many times, however, the historical data contains
only a few water quality parameters collected for treatment optimization or regulation
requirements. In most cases, baseline monitoring will need to be extended to a host of
other water quality parameters collected at locations that are not traditionally common
for routine sampling. Additional tests on distribution system pipes, such as XRD and SEM
analysis, can provide additional information related to the physical makeup of the pipe
scale surface but are not required as a part of a routine monitoring program. These meth-
ods are discussed in greater detail in chapter 5.
Monitoring Parameters
Routine water quality monitoring is useful for ensuring the effectiveness of, as well as
identifying the potential cause of, upsets in corrosion control treatment. Table 3-3 provides
a summary of recommended water quality monitoring parameters. Level 1, or baseline,
M58 book.indb 49 11/17/2010 4:24:50 PM

parameters should be monitored at routine intervals at the entry point and throughout
the distribution system, for example, daily and monthly, respectively. Level 2 parameters
are supplemental and may be measured in conjunction with Level 1 parameters if time
and budgets allow, or Level 2 may be measured for the purposes of verifying that optimal
conditions for corrosion control effectiveness have not been compromised. It may be help-
ful to first establish baseline concentrations of Level 2 parameters before deciding only to
monitor for these parameters on an intermittent basis. Level 3 parameters may be used to
pinpoint the probable cause of corrosion problems after the corrosion occurs and do not
need to be monitored on a regular basis.
The collection of any supplemental tap water samples should be carefully reviewed
with the state or federal primacy agency in advance, in order to appropriately distinguish
between (1) samples collected for LCR compliance and action level comparison and (2)
those efforts designed for the purposes of maintaining effective corrosion control treat-
ment. Utilities should also review their public communication policies and decide how
results of supplemental sampling will be communicated to those customers and others
that participate in the study.
Frequency of Monitoring
For optimal corrosion control treatment, USEPA requires utilities to monitor only a few
parameters, for example, pH, alkalinity, hardness, and phosphate (if an inhibitor is added)
at the entry point once every 2 weeks and twice every 6 months at locations in the distribu-
tion system. Reduced monitoring decreases distribution sampling considerably.
Compliance monitoring alone is insufficient for adequately characterizing system
conditions and monitoring corrosion control effectiveness. It is recommended that utili-
ties develop and implement an operational monitoring program to assist in optimizing
corrosion control treatment. Be certain to work out the arrangements of the operational
monitoring program with the appropriate primacy agency to understand what portion, if
any, of operational data may be reportable.
Table 3-3 Suggested water quality monitoring parameters

Level 1 (Baseline) Level 2 (Supplemental) Level 3 (Diagnostic)
pH Temperature Bacterial speciation
Total alkalinity Hardness (total and calcium) Sulfide
DIC (measured) Conductivity Nitrite/nitrate*
Phosphate (total and ortho-, if used) Dissolved oxygen Manganese
Disinfectant residual ORP
Ammonia (free and total)* Lead (total and dissolved)§
Chloride Copper (total and dissolved)§
Sulfate Iron (total and dissolved)
Aluminum† HPC
Iron (total and dissolved)‡ Color (apparent)
TOC
Turbidity
* Systems practicing chloramination only.
† System practicing alum or powdered activated carbon coagulation or those systems with high source water aluminum
concentrations.
‡ Systems utilizing ferric salts for coagulation, systems with a high percentage of iron mains, and those with high source water iron
concentrations.
§ Lead and copper samples may be required to be reported by the state or federal primacy agency. Discuss the implications of
additional lead and copper sampling prior to sample collection and analysis.
M58 book.indb 50 11/17/2010 4:24:50 PM

Entry-point readings should be taken more often than required by LCR compli-
ance monitoring, preferably on a daily basis. For systems with multiple entry points (e.g.,
groundwater systems with multiple wellfields), a program under which each entry point
is monitored routinely is still recommended; however, daily monitoring may not be practi-
cal or economical. In such cases, weekly monitoring may be more practical for several
reasons. The importance of entry-point monitoring is to ensure water quality entering the
distribution is consistent and stable. Small changes in source water pH, alkalinity, and DIC
may impact corrosion control treatment (i.e., water quality parameters) leaving the plant.
Increased distribution monitoring frequencies can determine areas where water may
stagnate or identify areas influenced by MIC.
After baseline levels have been established, monitoring of Level 2 and Level 3 param-
eters may be substantially reduced. However, monitoring frequencies should increase
whenever there is a change in source water, treatment, or distribution practices that
impacts water quality or physical factors that influence corrosion. For example, increased
monitoring of tap lead concentrations (total and dissolved) is recommended following a
change in pH or phosphate treatment. Similarly, a utility might conduct ORP, lead, and
nitrification monitoring before and after a change from chlorine to chloramine.
Monitoring Locations
Baseline water quality parameters should be monitored at the source, entry point, and
select locations in the distribution system. In addition to routine baseline monitoring, a
period of more frequent and targeted monitoring in the distribution system may be the
only practical method of gathering data on in situ materials and conditions. Such a tar-
geted monitoring program would include selected Level 2 and Level 3 parameters.
Required LCR monitoring targets single-family homes containing copper piping with
lead solder installed after 1982. When insufficient single-family homes are available to meet
this requirement, other buildings, including multifamily dwellings, may be sampled. It has
been demonstrated, however, that lead and copper problems may also occur in new homes
and plumbing compared with older homes and plumbing. Kimbrough (2007) found that
brass corrosion can be a major contributor to tap lead and copper concentrations, even in
homes with all plastic plumbing. Cantor (2000) demonstrated that new copper plumbing
may, in fact, be more susceptible to corrosion than older copper plumbing. Homes with
new or replaced meters and/or lead service lines may also be more susceptible to lead and
copper problems. Meter and service line replacement may result in particulate lead and
copper originating from the service lines (Sandvig et al. 2008). Partial replacement of lead
service lines may result in galvanic corrosion, in which the lead pipe itself becomes the
sacrificial anode. It is therefore critical to establish representative metal release rates in
the distribution system for all of these conditions.
For the reasons previously described, utilities may want to consider a supplemen-
tal operational tap water testing program. Such a program should carefully consider the
following:
• Sampling locations may need to be expanded to include conditions that represent
homes with new copper plumbing, brass faucets, new or replaced meters, and/or
replaced or partially replaced lead service lines. Ideally these new locations would
be homes or other building locations under the utility’s direct control.
• Utilization of TCR monitoring locations to assess the stability of distributed water
quality and possible impacts on corrosion may also be useful. Utilities should
review the Level 1, Level 2, and Level 3 parameters previously described to deter-
mine the most appropriate parameters for their system.
M58 book.indb 51 11/17/2010 4:24:50 PM

• Sample collection from some homes should be done in multiple 1-L samples: for
example, first liter, second liter, third liter, initial temperature change, 5 min. This
practice is known as profiling. Compliance with the LCR is based on the first-draw
sample (i.e., first liter of water flushed from the tap) based on the assumption that
the first-draw sample represents the worst-case sample. Because of the time and
expense involved, utilities may want to consider profiling only in certain situa-
tions, e.g., following lead service line replacement or in the event a change in tap
water lead or copper concentrations is observed. Chapter 5 discusses profiling in
greater detail.
Monitoring Program Example

Consider the following case study. A surface water system with moderate source water
alkalinity (60–100 mg/L as CaCO3) practices ferric chloride coagulation, flocculation, con-
ventional sedimentation, and dual media filtration. The system adds fluoride and uses free
chlorine for both primary and secondary disinfection. The distribution system includes a
substantial percentage of unlined cast-iron mains, and there are still a number of active
lead service lines. The system adds an orthophosphate–polyphosphate blend (70 percent
orthophosphate, 30 percent polyphosphate) at a dose of 1 mg P/L to minimize red water
and reduce tap lead concentrations.
Under current regulatory requirements, this system would be required to measure
the pH, alkalinity, and orthophosphate residual leaving the plant every other week. These
same parameters are required to be measured in the distribution system but only twice
every 6 months. Changes in raw water quality, such as pH, alkalinity, and turbidity, can
affect pH, alkalinity, and phosphate concentrations leaving the plant—particularly in sur-
face water systems. For this reason, this system chooses to monitor daily at the distribu-
tion entry point.
In addition to increased monitoring frequency at the system entry, this system has cho-
sen to increase the number of monitored parameters. Blended phosphates usually require
a tight pH range for optimizing corrosion control and iron sequestration at the same time.
If the pH fluctuates, polyphosphates that are bound to the iron can break down quickly to
orthophosphate and release iron as a ferrous hydroxide (Fe+2). Polyphosphate breakdown
can also happen naturally in areas of long detention time such as at the ends of large distri-
bution systems. If an oxidant such as chlorine is present, it will oxidize the iron into a ferric
precipitate, which could potentially cause discoloration at the customer’s tap. Therefore,
to optimize sequestration and corrosion control in the distribution system, monitoring for
total and dissolved iron, as well as total phosphate and orthophosphate, is practiced.
Because of the significant number of unlined cast-iron mains in the distribution sys-
tem and the fact that this system uses ferric chloride as a coagulant, this system also mon-
itors total and dissolved iron at the distribution entry point and selected TCR monitoring
locations (in areas where unlined cast-iron mains exist). Monitoring for these parameters
allows this system to trace red water events to treatment upsets or changes in the distribu-
tion system.
Historically, this system has experienced localized red water incidents in remote
areas of the distribution system. These areas are frequently characterized by high water
age and low chlorine residual. To guard against the possibility of MIC, this system has
opted to include HPC–R2A as a part of routine disinfectant residual monitoring.
M58 book.indb 52 11/17/2010 4:24:50 PM

ASSESSING THE CAUSE OF INTERNAL CORROSION

AND METALS RELEASE______________________________________
The purpose of this section is to provide guidance regarding how to utilize water quality and
other data to determine the most likely cause of internal corrosion and metals release in a
distribution system. As discussed in chapter 2, many mechanisms may result in increased
metals concentrations in drinking water distribution systems and home plumbing systems—
various types of uniform corrosion, nonuniform corrosion, and noncorrosion mechanisms.
For convenience in diagnosing a corrosion issue, these mechanisms are reclassified here as
chemically influenced, microbially influenced, and physically influenced.
Being able to quickly and accurately determine the cause of a corrosion-related inci-
dent—for example, red water or increases in tap lead concentrations—first requires that
baseline water quality be established. This baseline may then be compared to subsequent
water quality monitoring results to help pinpoint the cause of the event. For example,
• Is there a correlation between distribution chlorine residual data and the occur-
rence of home plumbing failures? Such a correlation could be an indicator that MIC
is occurring.
• What was the impact of a change in coagulant from alum to ferric chloride on the
finished water chloride-to-sulfate ratio?
• What impact did the addition of nanofiltration have on finished water pH and
alkalinity?
• Is there a need to modify the current corrosion control strategy?
Not only can water quality data help to determine the cause of a corrosion-related
event, but also it can be used to eliminate other possibilities. That is, if there were no sig-
nificant changes in key corrosion-related water quality parameters, there may have been
physical factors that led to the event. For example, if there were no changes in water
quality in an area of the distribution system that recently experienced an unprecedented
number of red water complaints, look to some of the other potential factors. Were there
system maintenance activities in the area that might have contributed to the event? Were
there changes in flow (velocity or direction) or system pressure that might have caused
shearing of existing pipe scale? Is there a storage tank nearby? Is that tank well mixed?
Is it possible that water discharged from that tank was of poor quality and contributed to
the event? Has the tank been cleaned recently? Is it possible sediment or other debris may
have come from the tank?
Table 3-4 provides a list of common corrosion-related water quality problems and the
potential causes of those problems in the context of three general categories. The list is by
no means exhaustive and can be broken down further into smaller categories as described
in chapter 2. The list is intended to be used as a guide in conjunction with the water quality
considerations discussed in this chapter when assessing water quality and other impacts
related to corrosion.
M58 book.indb 53 11/17/2010 4:24:50 PM

Table 3-4 Assessment of common corrosion-related water quality problems
Indicators That the Symptom May Be . . .

Symptom Chemically Influenced Microbially Influenced Physically Influenced
Red water • Shift in dissolved/ • Loss or reduction in • Changes in flow
particulate iron speciation disinfectant residual • Changes in system
but no change in total iron • Increase in HPC–R2A pressure
concentration • Localized shifts in pH but • System maintenance
• Loss or reduction in not systemwide • Release of sediments from
disinfectant residual • Increases in distribution distribution storage tanks
• Change in ORP nitrite/nitrate • Release of lower quality
• Shift in pH concentrations* water from poorly mixed
• Change in distributed • Higher or lower than storage tanks
water alkalinity normal free ammonia • Distribution storage tank
• Treatment process upsets concentrations* cleaning
Increased • Recent change in • Loss or reduction in • Increase in particulate

tap lead distribution system disinfectant residual lead concentration
concentrations disinfectant • Increase in HPC–R2A but no corresponding
• Change in ORP • Localized shifts in pH but increase in dissolved lead
• Drop in finished water pH not systemwide concentrations
• Shift in pH to outside • Increases in distribution • Recent lead service line
optimal range for corrosion nitrite/nitrate replacement
control chemical concentrations* • Installation of new brass
• Change in distributed • Higher or lower than plumbing fixture
water alkalinity normal free ammonia • Long periods of stagnation
• Change in finished water concentrations*
chloride-to-sulfate ratio
• Poor buffer intensity
Increased • Recent change in • Loss or reduction in • Increase in particulate
tap copper secondary disinfectant disinfectant residual copper concentration, but
concentrations • Change in ORP • Increase in HPC–R2A no corresponding increase
• Drop in finished water pH • Localized shifts in pH but in dissolved copper
• Shift in pH to outside not systemwide concentrations
optimal range for corrosion • Increases in distribution • Long periods of stagnation
control chemical nitrite/nitrate
• Change in distributed concentrations*
water alkalinity • Higher or lower than
normal free ammonia
concentrations*
• Presence of sulfide
Pinhole leaks in • pH below 7.2 • Loss or reduction in • Stagnation
copper plumbing • High sulfate relative to disinfectant residual • Sediment or debris in
chloride and bicarbonate • Increase in HPC–R2A plumbing
• Treatment upsets resulting • Localized shifts in pH but • High velocity (erosion
in increased residual not systemwide corrosion)
aluminum concentration in • Increases in distribution • Improper fluxing (leads to
finished water nitrite/nitrate “tracks” on inside of pipe)
concentrations*
• Higher or lower than
normal free ammonia
concentrations*
• Presence of sulfide
Table continued next page.
M58 book.indb 54 11/17/2010 4:24:51 PM

Table 3-4 Assessment of common corrosion-related water quality problems (continued)
Indicators That the Symptom May Be . . .

Symptom Chemically Influenced Microbially Influenced Physically Influenced
Blue water • Drop in pH below 7.2 in • Loss or reduction in • Stagnation
soft waters disinfectant residual
• Rise in pH above 8.0 in • Increase in HPC–R2A
soft waters • Localized shifts in pH but
• Drop in finished water not systemwide
alkalinity • Increases in distribution
nitrite/nitrate
concentrations*
• Higher or lower than
normal free ammonia
concentrations*
Pink water • pH shift • Pink water is not likely to • Changes in flow
• Treatment upsets resulting be the result of MIC • Changes in system
in increased MnO4– pressure
concentrations entering • System maintenance
distribution system • Release of sediments from
distribution storage tanks
• Release of lower quality
water from poorly mixed
storage tanks
• Distribution storage tank
cleaning
Black or yellow • Drop in finished water ORP • Black or yellow water is • Changes in flow
water • Treatment upsets resulting not likely to be the result • Changes in system
in reduction of MnO2 of MIC pressure
• System maintenance
• Release of sediments from
distribution storage tanks
• Release of lower quality
water from poorly mixed
storage tanks
• Distribution storage tank
cleaning
* Systems adding chloramine.
SUMMARY_________________________________________________
This chapter has discussed water quality factors and other factors that influence corro-
sion and has provided the basis for a corrosion control monitoring program. It also has
provided a framework by which utilities may assess and determine the cause of corrosion-
related water quality problems in the context of three general categories of corrosion—
chemically influenced, microbially influenced, and physically influenced corrosion.
Understanding the carbonate balance and its impact on corrosion and corrosion con-
trol treatment is critical to effective corrosion control. Other water quality parameters,
such as oxidants, ORP, anions, and biostability, may also have a significant impact on
metals release. In fact, MIC is often overlooked as a source of corrosion-related water
quality problems, and recognition of the importance of biostability and its role in corro-
sion may help to eliminate many of the corrosion-related water quality problems utilities
experience.
M58 book.indb 55 11/17/2010 4:24:51 PM

The level of monitoring required by the LCR is insufficient for maintaining effective
corrosion control treatment. More frequent monitoring of LCR-required parameters and
other parameters at the source, entry point, and locations throughout the distribution sys-
tem is recommended. It is also necessary to establish baseline water quality to effectively
identify and implement corrosion control treatment and subsequently identify the cause
of any potential corrosion-related water quality problems. Utilities may consider the use
of compliance monitoring results to establish baseline water quality and effectiveness of
the corrosion control treatment, and then develop a supplemental operational monitoring
program to better manage and control the effectiveness of the program.
Lack of an adequate monitoring program will make it difficult to determine the cause
of corrosion-related water quality problems. Identifying the cause of a corrosion-related
water quality problem requires that one understands how three different categories of cor-
rosion look and how they might possibly be linked to that problem.
ADDITIONAL READING______________________________________
Readers may wish to review more information in the specific areas of
• The carbonate balance, more specifically pH, alkalinity, and DIC, and the role of
these parameters in corrosion-related water quality problems and effective corro-
sion control
• The role of other water quality factors on corrosion and the ability to effectively
control metals release
• The importance of biostability and the role of MIC in corrosion-related water qual-
ity problems
References for this chapter give detailed information on these areas and the other
topics discussed. Readers are encouraged to follow up on the referenced material
that is publicly available to obtain keener insight into this discussion.
In addition to the references, the following resources provide substantial discus-
sions regarding these topics and will be extremely valuable to readers in devel-
oping an understanding of the factors that influence corrosion, implementing an
effective corrosion monitoring program, and assessing the cause of future corro-
sion-related water quality issues.
• Economic and Engineering Services and Illinois State Water Survey, Lead Control
Strategies (Denver, Colo.: AwwaRF and AWWA 1990)
• AwwaRF and TZW, Internal Corrosion of Water Distribution Systems, 2nd edi-
tion (Denver, Colo.: AwwaRF and AWWA 1996)
• Marc Edwards, Travis E. Meyer, John Rehring, John Ferguson, Gregory Korshin,
and Samuel Perry. Role of Inorganic Anions, Natural Organic Matter and Water
Treatment Process in Copper Corrosion (Denver, Colo.: AwwaRF and AWWA
1996)
• Gregory V. Korshin, John F. Ferguson, Alice N. Lancaster, and Hao Wu, Corrosion
and Metal Release for Lead Containing Materials: Influence of NOM (Denver,
Colo.: AwwaRF and AWWA 1999)
• G.J. Kirmeyer, G. Pierson, J. Clement, A. Sandvig, V. Snoeyink, W. Kriven, and
A. Camper. Distribution System Water Quality Changes Following Corrosion
Control Strategies. (Denver, Colo.: AwwaRF and AWWA 2000)
M58 book.indb 56 11/17/2010 4:24:51 PM

• M. Edwards, L.S. McNeill, and T.R. Holm. Role of Phosphate Inhibitors in Mitigat-
ing Lead and Copper Corrosion (Denver, Colo.: AwwaRF and AWWA 2001)
• G. Kirmeyer, B. Murphy, and A. Sandvig. Post Optimization of Lead and Copper
Control Monitoring Strategies (Denver, Colo.: AwwaRF and AWWA 2004)
• Gregory Kirmeyer, Kathy Martel, Gretchen Thompson, Lori Radder, Wyndi Klement,
Mark LeChevallier, Helene Baribeau, and Andrea Flores. Optimizing Chloramine
Treatment, 2nd edition (Denver, Colo.: AwwaRF and AWWA 2004)
• T. Case. Distribution System Corrosion and the Lead and Copper Rule: An Over-
view of AwwaRF Research (Denver, Colo.: AwwaRF and AWWA 2004)
REFERENCES________________________________________________
Abernathy, C.G., and Camper, A.K. 1998. The Broo, A.E., B. Berghult, and T. Hedburg. 1999.
Effect of Phosphorous-Based Corrosion Drinking Water Distribution—the Effect
Inhibitors and Low Disinfectant Residuals of Natural Organic Matter (NOM) on the
on Distribution System Biofilms. In Proc. Corrosion of Iron and Copper. Water Sci.
of the AWWA Water Quality Technology Tech., 40(9):17–24.
Conference. Denver, Colo.: AWWA. Burlingame, G.A., D.A. Lytle, and V.L. Snoeyink.
American Water Works Association (AWWA). 2006. Why Red Water? Understanding Iron
2005. Managing Change and Unintended Release in Distribution Systems. Opflow,
Consequences: Lead and Copper Rule 32(12):12–16.
Corrosion Control Treatment. Denver, Burlingame, G.A., M.R. Schock, and M.A.
Colo.: AWWA. Edwards. 2006. Knowing Chemistry
Arweiler, S., Y.K. Cohen, and C.G. Abernathy. Can Help Get the Lead Out. Opflow,
2003. Polyphosphates—the Solution to 32(9):24–26.
Distribution System Low Residuals, Bio- Campbell, H.S., and M.E.D. Turner. 1983. The
film and Pipe Corrosion—Part II. In Proc. Influence of Trace Organics on Scale For-
of the AWWA Water Quality Technology mation and Corrosion. Jour. Inst. Water
Conference. Denver, Colo.: AWWA. Eng. and Science, 4:55.
Benjamin, M.M., H. Sontheimer, and P. Leroy. Cantor, A.F. 2000. Use of Polyphosphate
1996. Corrosion of Iron and Steel. In Inter- in Corrosion Control. Jour. AWWA,
nal Corrosion of Water Distribution Sys- 92(2):95–102.
tems. 2nd ed. Denver, Colo.: AwwaRF and Cantor, A.F. 2003. Effect of Chlorine on Cor-
AWWA. rosion in Drinking Water Systems. Jour.
Boissonneault, P. 1994. Dissolved Inorganic AWWA, 95(5):112–123.
Carbon—a Key to Corrosion Control in Cantor, A.F. 2006. Diagnosing Corrosion Prob-
Low Alkalinity Waters. In Proc. of the lems Through Differentiation of Metal
AWWA Water Quality Technology Confer- Fractions. Jour. AWWA, 98(1):117.
ence. Denver, Colo.: AWWA. Cantor, A.F., J.B. Bushman, and M.S. Glodoski.
Boulay, N. and M. Edwards. 2001. Role of 2003. A New Awareness of Copper Pipe
Temperature, Chlorine, and Organic Failures in Water Distribution Systems. In
Matter in Copper Corrosion By-product Proc. of the AWWA Water Quality Tech-
Release in Soft Water. Wat. Research, 35 nology Conference. Denver, Colo.: AWWA.
(3):683–690. Cantor, A.F., J.B. Bushman, M.S. Glodoski, E.
Boyd, G.R., P. Shetty, A.M. Sandvig, and G.L. Pier- Kiefer, R. Bersch, and H. Wallenkamp.
son. 2004. Pb in Tap Water Following Simu- 2006. Copper Pipe Failure by Microbio-
lated Partial Lead Pipe Replacements. Jour. logically Influenced Corrosion. Materials
Env. Engr., 130(10):1188–1197. Performance (June 2006):38–41.
Bremer, P.J., B.J. Webster, and D.B. Wells. 2001. Carlson, K., S. Via, B. Bellamy, and M. Carlson.
Biocorrosion of Copper in Potable Water. 2000. Secondary Effects of Enhanced
Jour. AWWA, 93(8):82–91. Coagulation and Softening. Jour. AWWA,
92(6):63–75.
M58 book.indb 57 11/17/2010 4:24:51 PM

Clement, J.A., and M.R. Schock. 1998. Buffer Ferguson, J.F., O. von Franque, and M.R. Schock.
Intensity: What Is It and Why It’s Critical 1996. Corrosion of Copper in Potable Water
for Controlling Distribution System Water Systems. In Internal Corrosion of Water
Quality. In Proc. of the AWWA Water Qual- Distribution Systems. 2nd ed. Denver,
ity Technology Conference. Denver, Colo.: Colo.: AwwaRF and AWWA.
AWWA. Fleming, K.K., G.W. Harrington, and D.R. Nogu-
Clesceri, N.L., and G.F. Lee. 1965. Hydrolysis of era. 2005. Nitrification Potential Curves: A
Condensed Phosphates—II: Sterile Envi- New Strategy for Nitrification Prevention.
ronment. Int. Jour. Air and Water Pollu- Jour. AWWA, 97(8):90.
tion, 9:743. Holm, T.R., and M. Edwards. 2003. Meta-
Dodrill, D.M., L. Didmi, and M. Edwards. 1996. phosphate Reversion in Laboratory and
Role of Inhibitors, Chloride to Sulfate Pipe-Rig Experiments. Jour. AWWA,
Ratios and Color in Lead and Copper Cor- 95(4):172–178.
rosion By-Product Release. In Proc. of the Holm, T.R., and Schock, M.R. 1991. Potential
AWWA Annual Conference and Exposi- Effects of Polyphosphate Products on
tion. Denver, Colo.: AWWA. Lead Solubility in Plumbing Systems.
Dodrill, D.M., and M. Edwards. 1995. Corrosion Jour. AWWA, 83(7):76–82.
Control on the Basis of Utility Experience. Horsley, M.B., B.W. Northup, W.J. O’Brien, and
Jour. AWWA, 87(7):74–85. L.L. Harms. 1998. Minimizing Iron Cor-
Economic and Engineering Services and Illi- rosion in Lime Softened Water. In Proc.
nois State Water Survey. 1990. Lead Con- of the AWWA Water Quality Technology
trol Strategies. Denver, Colo.: AwwaRF Conference. Denver, Colo.: AWWA.
and AWWA. Imran, S.A., J.D. Dietz, G. Mutoti, J.S. Taylor,
Edwards, M. 2004. Letter report to Stafford A.A. Randall, and C.D. Cooper. 2005. Red
County, Va., July 22, 2004. Water Release in Drinking Water Distribu-
Edwards, M., and A. Dudi. 2004. Role of Chlo- tion Systems. Jour. AWWA, 97(9):93–100.
rine and Chloramine in Corrosion of Lead- Ives, D.J., and R.W. Rawson. 1962. Copper Cor-
Bearing Plumbing Materials. Jour. AWWA, rosion IV: The Effects of Saline Additions.
96(10):69–81. Jour. Electrochemical Soc., 109(6):462.
Edwards, M., J.F. Ferguson, and S.H. Reiber. Jacobs, S., S. Reiber, and M. Edwards. 1998.
1994. The Pitting Corrosion of Copper. Sulfide-Induced Copper Corrosion. Jour.
Jour. AWWA, 86(7):74–90. AWWA, 90(7):62–73.
Edwards, M., S. Jacobs, and D. Dodrill. 1999. Kimbrough, D.E. 2007. Brass Corrosion as a
Desktop Guidance for Mitigating Pb and Source of Lead and Copper in Traditional
Cu Corrosion By-Products. Jour. AWWA, and All-Plastic Distribution Systems.
91(5):66–77. Jour. AWWA, 99(8):70–76.
Edwards, M., S. Jacobs, and R.J. Taylor. 2000. Kirchner, G. 2009. Clintonville School Water OK
The Blue Water Phenomenon. Jour. to Drink. Shawano Leader, Feb. 25, 2009.
AWWA, 92(7):72–82. Korshin, G.V., J.F. Ferguson, and A.N. Lancaster.
Edwards, M., T. Meyer, and J. Rehring. 1994. 2005. Influence of Natural Organic Matter
Effect of Selected Anions on Copper Cor- on the Morphology of Corroding Lead Sur-
rosion Rates. Jour. AWWA, 86(12):73–81. faces and the Behavior of Lead-Containing
Edwards, M., and L.S. McNeill. 2002. Effect of Particles. Wat. Research, 39(5):811–818.
Phosphate Inhibitors on Lead Release Landers, J. 2006. Getting the Lead Out. Public
From Pipes. Jour. AWWA, 94(1):79–90. Works, 137(4):30–32.
Edwards, M. and S. Reiber. 1997. A General Larson, T.E. 1966. Chemical Control of Corro-
Framework for Corrosion Control Based sion. Jour. AWWA, 58(3):354.
on Utility Experience. Denver, Colo.: Lauer, W.C., and S.R. Lohman. 1994. Non-
AwwaRF and AWWA. Calcium Carbonate Protective Film Low-
Edwards, M., M.R. Schock, and T.E. Meyer. ers Lead Levels. In Proc. of the AWWA
1996. Alkalinity, pH, and Copper Corro- Water Quality Technology Conference.
sion By-Product Release. Jour. AWWA, Denver, Colo.: AWWA.
88(3):81–94.
Escobar, I.C., and A.A. Randall. 2001. Case
Study: Ozonation and Distribution System
Biostability. Jour. AWWA, 93(10):77.
M58 book.indb 58 11/17/2010 4:24:51 PM

LeChavallier, M.W., C.D. Lowry, R.G. Lee, and Reiber, S.H., J.F. Ferguson, and M.M. Benjamin.
D.L. Gibbon. 1993. Examining the Rela- 1987. Corrosion Monitoring and Control
tionship Between Iron Corrosion and the in the Pacific Northwest. Jour. AWWA,
Disinfection of Biofilm Bacteria. Jour. 79(2):71–74.
AWWA, 85(7):111–123. Renner, R. 2006. Experiment Confirms
Lee, S.H., J.T. O’Conner, and S.K. Banerji. 1980. Chloramine’s Effect on Lead in Drinking
Biologically Mediated Corrosion and Its Water. Environ. Sci. Technol. A-Pages,
Effects on Water Quality in the Distribu- 40(10):3129–3130.
tion System. Jour. AWWA, 72(11):636. Rezania, L. 2004. Optimizing Phosphate Treat-
Lytle, D.A., P. Sarin, and V.L. Snoeyink. 2003. ment to Minimize Lead/Copper Sea-
The Effect of Chloride and Orthophos- sonal Variations. Waterline, Minnesota
phate on the Release of Iron From Drink- Department of Health, 12(3):2 (winter
ing Water Distribution System Cast Iron 2004–2005).
Pipe. In Proc. of the AWWA Water Qual- Robinson, R.B., R.A. Minear, and J.M. Holden.
ity Technology Conference. Denver, Colo.: 1987. Effects of Several Ions on Iron Treat-
AWWA. ment by Sodium Silicate and Hypochlo-
Lytle, D.A., and M.R. Schock. 2005. Formation rite. Jour. AWWA, 79(7):116.
of Pb(IV) Oxides in Chlorinated Water. Rushing, J.C., and M. Edwards. 2002. Effect of
Jour. AWWA, 97(11):102–114. Aluminum Solids and Chlorine on Cold
Maynard, B., and D. Mast. 2006. Composition of Water Pitting of Copper. In Proc. of the
Interior Scales on Lead Source Materials. AWWA Water Quality Technology Confer-
In Proc. of the AWWA Water Quality Tech- ence. Denver, Colo.: AWWA.
nology Conference. Denver, Colo.: AWWA. Sandvig, A., P. Kwan, G. Kirmeyer, B. Maynard,
McNeill, L.S., and M. Edwards. 2000. Phosphate D. Mast, R. Rhodes Trussell, S. Trussell,
Inhibitors and Red Water in Stagnant A. Cantor, and A. Prescott. 2008. Con-
Pipes. Jour. Env. Engr., 126(12):1096. tribution of Service Line and Plumb-
McNeill, L.S., and M. Edwards. 2001. Iron Pipe ing Fixtures to Lead and Copper Rule
Corrosion in Distribution Systems. Jour. Compliance Issues. Denver, Colo.: Water
AWWA, 93(7):88–100. Research Foundation.
McNeill, L.S., and M. Edwards. 2002. Phos- Scardina, P.W. 2006. Copper Pinhole Leaks. In
phate Inhibitor Use at U.S. Utilities. Jour. Proc. of the AWWA Annual Conference
AWWA, 94(7):57–63. and Exposition. Denver, Colo.: AWWA.
Obrecht, M.F., and L.L. Quill. 1960. How Temper- Schock, M.R. 1999. Internal Corrosion and
ature, Treatment, and Velocity of Potable Deposition Control. In Water Quality and
Water Affect Corrosion of Copper and Its Treatment. 5th ed. New York: McGraw-
Alloys; Tests Show Effects of Water Qual- Hill.
ity at Various Temperatures, Velocities. Schock, M.R., J. Clement, D.A. Lytle, A.M.
Heating, Piping, Air Cond., 32(5):105. Sandvig, and S.M. Harmon. 1998. Replac-
Obrecht, M.F., and L.L. Quill. 1961. How Tem- ing Polyphosphate With Silicate to Solve
perature, Velocity of Potable Water Affect Problems With Lead, Copper and Source
Corrosion of Copper and Its Alloys. Heat- Water Iron. In Proc. of the AWWA Water
ing, Piping, and Air Cond., 33(4):129. Quality Technology Conference. Denver,
Pisigan, R.A., and J.E. Singley. 1987. Influence Colo.: AWWA.
of Buffer Capacity, Chlorine Residual, and Schock, M.R., T.L. Geske, M.K. DeSantis, and
Flow Rate on Corrosion of Mild Steel and R.C. Copeland. 2006. Scale Analysis for
Copper. Jour. AWWA, 79(2):62–70. Madison, WI, Samples. Cincinnati, Ohio:
Reiber, S. 2004. Disinfection Byproducts vs. USEPA.
Corrosion: A Case Study on the DC WASA Schock, M.R., and R. Giani, R. 2005. Oxidant/
Lead Experience. Presentation to USEPA Disinfectant Chemistry and Impacts on
Technical Expert Working Group. April Lead Corrosion. In Proc. of the AWWA
2004. Washington, D.C. Water Quality Technology Conference.
Reiber, S. 2006. Corrosion, Red Water, and Dis- Denver, Colo.: AWWA.
tribution System Water Quality. In Proc. of
the AWWA Annual Conference and Expo-
sition. Denver, Colo.: AWWA.
M58 book.indb 59 11/17/2010 4:24:51 PM

Schock, M.R., and D.A. Lytle. 1995. Control of Stericker, W. 1945. Protection of Small Water
Copper Corrosion of Household Plumb- Systems From Corrosion. Industrial
ing Materials. In Abstract Proc. of the 21st Engineering Chemistry, 37:716.
Annual USEPA RREL Research Sympo- Stone, A., D. Spyridakis, M. Benjamin, J. Fergu-
sium. Cincinnati, Ohio. son, S. Reiber, and S. Osterhus. 1987. The
Schock, M.R., D.A. Lytle, A.M. Sandvig, J. Clem- Effects of Short Term Changes in Water
ent, and S.M. Harmon. 2005. Replacing Quality on Copper and Zinc Corrosion
Polyphosphate With Silicate to Solve Lead, Rates. Jour. AWWA, 79(2):75–82.
Copper, and Source Water Iron Problems. Stumm, W. 1960. Investigation of the Corro-
Jour. AWWA, 97(11):84–93. sive Behavior of Waters. ASCE Jour. San.
Schock, M.R., I. Wagner, and R.J. Oliphant. Eng. Division, 86(SA6):27.
1996. Corrosion and Solubility of Lead in Triantafyllidou, S., and M. Edwards. 2007. Role
Drinking Water. In Internal Corrosion of Chloride to Sulfate Mass Ratio in Lead
of Water Distribution Systems. 2nd ed. Leaching From Soldered Joints and Brass.
Denver, Colo.: AwwaRF and AWWA. Jour. AWWA, 99(7):96-109.
Singley, J.E. 1981. The Search for a Corrosion Trussell, R.R., and I. Wagner. 1996. Corrosion
Index. Jour. AWWA, 73(11):579–582. of Galvanized Pipe. In Internal Corro-
Singley, J.E. 1994. Electrochemical Nature sion of Water Distribution Systems. 2nd
of Lead Contamination. Jour. AWWA. ed. Denver, Colo.: AwwaRF and AWWA.
86(7):91–96. Tuovinen, O.H., K.S. Button, A. Vuorinen,
Smith, C.D. 2006. Monitoring for Nitrification L. Carlson, D.M. Mair, and L.A. Yut.
Prevention and Control. In M56: Funda- 1980. Bacterial, Chemical, and Miner-
mentals and Control of Nitrification in alogical Characteristics of Tubercles
Chloraminated Drinking Water Distri- in Distribution Pipelines. Jour. AWWA,
bution Systems. Denver, Colo.: AWWA. 72(11):625–626.
Smith, S.E., T. Ta, D.M. Holt, A. Delanoue, and J. US Environmental Protection Agency (USEPA).
Colbourne. 1998. Minimising Red Water in 2007. Elevated Lead in DC Drinking
Drinking Water Distribution Systems. In Water. EPA 815-R-07-021. Washington,
Proc. of the AWWA Water Quality Tech- D.C.: USEPA, Office of Water.
nology Conference. Denver, Colo.: AWWA. Van Der Merwe, S.W. 1988. The Effect of Water
Snoeyink, V.L., P. Sarin, and D.A. Lytle. 2003. Quality Variables on the Corrosive Behav-
Iron Release and Colored Water Forma- iour of Water Coagulated With a Cationic
tion From Iron Scales. In Proc. of the Polyelectrolye and With Lime/Activated
AWWA Annual Conference and Exposi- Silica. Water Supply, 6(4):SS2.
tion. Denver, Colo.: AWWA. Van der Kooij, D. 1992. Assimilable Organic
Snoeyink, V.L., and I. Wagner. 1996. Principles Carbon as an Indicator of Bacterial
of Corrosion of Water Distribution Sys- Regrowth. Jour. AWWA, 84(2):57.
tems. In Internal Corrosion of Water Dis- Vik, E.A., R.A, Ryder, I. Wagner, and J.F. Fergu-
tribution Systems. 2nd ed. Denver, Colo.: son. 1996. Mitigation of Corrosion Effects.
AwwaRF and AWWA. In Internal Control of Water Distribu-
Sontheimer, H., W. Kolle, and R. Rudek. 1979. tion Systems. New York: McGraw-Hill.
Aufgaben und Methoden der Wasserche- Williams, S.M. 1990. The Use of Sodium Silicate
mie—Dargestellt an der Entwicklung and Sodium Polyphosphate to Control
erkenntnisse zur Bildung von Korrosions- Water Problems. Water Supply, 8:195.
shutzschichten auf Metallen. Vom Wasser, Zhang, M., M.J. Semmens, D. Schuler, and R.M.
52:1. Hozalski. 2002. Biostability and Microbio-
Sontheimer, H., W. Kolle, and V.L. Snoeyink. logical Quality in a Chloraminated Distri-
1981. The Siderite Model of the Forma- bution System. Jour. AWWA, 94(9):112.
tion of Corrosion-Resistant Scales. Jour.
AWWA, 73(11):572.
Sorg, T.J., M.R. Schock, and D.A. Lytle. 1999.
Ion Exchange Softening: Effects on
Metals Concentrations. Jour. AWWA,
91(8):85–97.
M58 book.indb 60 11/17/2010 4:24:51 PM

AWWA MANUAL M58
Chapter 4
Corrosion Control
Techniques
Elizabeth Turner
City of Dallas, Texas
Richard Giani
DC Water
INTRODUCTION_____________________________________________
The implementation of corrosion control can be accomplished in eight steps.
1. Develop understanding of factors affecting internal corrosion
2. Determine extent and magnitude of corrosion
3. Determine the possible causes of corrosion
4. Assess corrosion control alternatives
5. Select a corrosion control strategy
6. Implement a corrosion control program
7. Monitor the effectiveness of the corrosion control program
8. Optimize the control program if necessary
Selection of a corrosion control technique may not be a one-time event. Any changes
in water treatment or source water will require a system to reevaluate corrosion con-
trol and repeat steps 1 through 6. Earlier chapters have discussed corrosion theory and
how to determine possible causes of corrosion. This chapter will cover the assessment
of corrosion control alternatives and some tools to help select a corrosion control strat-
egy (Table 4-1). The complete elimination of corrosion is almost impossible. Technologies
exist to reduce or inhibit corrosion, but the feasibility of each technology will vary from
system to system. The success of a particular corrosion control treatment is dependent on
the specific water quality and piping of an individual system.
61
M58 book.indb 61 11/17/2010 4:24:52 PM


• The piping materials available for use in a distribution system and premise plumbing include iron, cement, steel,
lead, copper, brass, and plastic. Selection of material will be based on water chemistry of the system and cost.
• The design of pipes and structures in the distribution system can be as important as the selection of materials.
• Routine system maintenance such as flushing, cleaning, and lining of mains can reduce corrosion and/or remove
the by-products of corrosion.
• Chemical treatment options include pH adjustment, dissolved inorganic carbon (DIC) adjustment, and
application of corrosion inhibitors.
• Techniques for controlling microbially influenced corrosion (MIC) include disinfection, control of water age,
flushing, and nutrient control.
• Utilities may find it necessary to reoptimize corrosion control treatment whenever changes are encountered in
the source water or when utilities are contemplating treatment changes such as changes of coagulant type or
secondary disinfectant.
Implementing and maintaining optimum corrosion control treatment require balanc-

ing the needs of conflicting water quality goals. For example, the optimum pH for coagu-
lation is not the optimum pH for orthophosphate used for corrosion control. The most
common techniques for controlling corrosion include
• Distribution system design considerations
• Coatings and linings
• Water quality modifications
• Corrosion inhibitors
DISTRIBUTION SYSTEM DESIGN CONSIDERATIONS_______________

A water utility can minimize corrosion by properly selecting distribution system materi-
als and having a good engineering design. The reality is, however, that most utilities are
forced to deal with existing distribution systems that were designed with little regard for
corrosion. Nonetheless, utilities can take advantage of system design for corrosion con-
trol during system upgrades, planned system materials replacement, and expansion of the
System and Pipe Materials

The piping materials available for use in a distribution system and premise plumbing include
iron, cement, steel, lead, copper, brass, and plastic. Water mains installed between the late
1800s and the 1920s are usually unlined cast iron. Ductile-iron pipe began to be installed
in the 1950s. Polyvinyl chloride (PVC) pipes were introduced in the 1970s followed by high
density polyethylene (HDP) in the 1990s. It is estimated that 20 percent of the water mains
in North America are lined with asbestos–cement or cement–mortar. Most of the ductile-
iron pipe and approximately 40 percent of the cast-iron mains are mortar-lined.
The selection of piping material will be based on water chemistry of the system and
cost. The mechanical properties of the piping material will also influence the selection of
materials. In addition, some jurisdictions regulate the components of the distribution system.
For example, Naperville, Ill., requires its mains to be cement-lined ductile-iron pipe, with a
minimum Class 52 thickness designation, and polyethylene encasement. It also requires that
service lines 2 in. (5.08 cm) or less (<2 in.) inside diameter must be copper pipe (Type K),
and all services 3 in. (7.62 cm) or greater (>3 in.) inside diameter must be Class 52 duc-
tile-iron pipe with cement lining. When selecting materials to replace old or broken mains,
M58 book.indb 62 11/17/2010 4:24:52 PM

Corrosion Control Techniques 63
Table 4-2 Corrosion properties of materials frequently used in water distribution systems
Primary Contaminant
Plumbing Material Corrosion Resistance From Pipe
Asbestos–cement, Good corrosion resistance; immune to electrolysis; Asbestos fibers, increase in pH,
concrete, and cement aggressive waters can leach calcium from the aluminum, and calcium. Trace
linings cement; polyphosphate sequestering agents can metals (cadmium, chromium,
deplete the calcium and substantially soften the barium, and aluminum) due to
pipe. presence in cement.
Brass Good overall resistance; different types of brass Lead, copper, zinc
respond differently to water chemistry; subject to
dezincification by waters of pH >8.3 with high ratio
of chloride-to-carbonate hardness.
Copper Good overall corrosion resistance; subject to Copper and possibly iron, zinc,
corrosive attack from high flow velocities, soft tin, antimony, arsenic, cadmium,
water, chlorine, dissolved oxygen, low pH, and high and lead from associated pipes
inorganic carbon levels (alkalinity). Subject to MIC. and solder.
May be prone to pitting failures.
Galvanized iron Subject to galvanic corrosion of zinc by aggressive Zinc and iron. Cadmium,
or steel waters, especially of low hardness; corrosion is chromium, barium, aluminum,
accelerated by contact with copper materials; and lead are possible due to
corrosion is accelerated at higher temperatures such impurities in the galvanization
as in hot water systems; corrosion is affected by process.
manufacturing process of the pipe and galvanized
coating.
Iron, unlined cast Can be subject to surface erosion by aggressive Iron. Resulting in red water,
or ductile waters and tuberculation in poorly buffered waters. complaints, and turbidity.
Lead Corrodes in soft water with pH <8 and in hard Lead
waters with high inorganic carbon levels and pH
< 7.5 or > 8.5.
Plastic Resistant to corrosion Some pipes contain metals
in plasticizers, notably lead.
Plasticizers are an emerging
contaminant issue.
Steel, mild Subject to uniform corrosion; affected primarily by Iron. Resulting in red water,
high dissolved oxygen and chlorine levels and poorly complaints, and turbidity.
buffered water.
putting new mains into service, or rehabilitating existing pipes, the utility should choose
a material that will not corrode in the environment (external) or under the water quality
conditions anticipated (internal). Table 4-2 will assist in the selection of piping materials.
Comparable materials should be used throughout the system. Galvanic corrosion
should be minimized by not having two materials with different electrical potentials, such
as copper and galvanized iron, come in direct contact. The placement of dielectric (insu-
lating) couplings between dissimilar metals is highly recommended.
Asbestos–cement pipe, cement–mortar linings, and concrete pipe. Asbestos–
cement pipe (see Figure 4-1) was once popular because of its corrosion resistance. Health
concerns from the potential release of asbestos fibers have limited the installation of
new asbestos–cement pipe (AwwaRF and TZW 1996). Cement pipe is generally resistant
to corrosion. Asbestos–cement pipe and prestressed concrete cylinder pipe behave very
similarly with the exception of the possible release of asbestos from asbestos–cement
pipes with deterioration of that pipe. The leaching of calcium and its complexes from
cement materials raises the pH, which can reduce the effectiveness of disinfectants, and
M58 book.indb 63 11/17/2010 4:24:52 PM

can cause unwanted precipitation of minerals, thus resulting in cloudy or turbid water and
poor taste.
The cement matrix can be quite complex. More than 100 compounds and phases
important to the chemistry of Portland cement and related cements have been described
and identified (Schock 1989). The main components of Portland cement are tricalcium
silicate, dicalcium silicate, and tricalcium aluminate plus small amounts of iron and
magnesium.
The corrosion resistance of mortar linings is dependent on the mortar density. Mixture
of cement and sand in a 1:1 ratio is more resistant to concrete corrosion than other ratios.
One of the best protective measures for all cement-based materials is calcium car-
bonate deposition, which can fill voids in the cement matrix and prevent leaching. Unfor-
tunately for asbestos–cement pipes, calcium carbonate deposition still leaves asbestos
fibers exposed and vulnerable to erosion. Iron coatings have been found to be beneficial
by preventing exposure of the fibers at the pipe surface and reducing the possibility of
release of asbestos fibers into the bulk water. However, they do not seal the cement pipe
matrix from dissolution. Iron coatings may also have their own corrosion-related prob-
lems. Manganese coatings can provide a similar protection to that provided by iron. Silica
may also be beneficial by assisting the formation and adsorption of iron colloids onto the
pipe surface.
Zinc orthophosphate, zinc sulfate, and zinc chloride have been used in the lab to
prevent softening of cement pipes. Strong sequestering agents such as polyphosphates
can attack the pipe by enhancing calcium, aluminum, iron, and magnesium from the
cement matrix.
Bronze and brasses. Bronze and brasses are primarily used in home plumbing fix-
tures and fittings (Figures 4-2 and 4-3). Bronze refers to a broad range of copper alloys,
usually with tin as the main additive but sometimes with other added elements such as
phosphorus, manganese, aluminum, or silicon. Brass is the term used for alloys of copper
and zinc. The amount of zinc may vary from 5 to 45 percent. Lead is often added to brass
to improve machinability. Valves, faucets, and water meters may be composed of brass,
bronze, or other materials. Most studies on corrosion of brass and bronze have focused on
dezincification. The corrosion process can contribute lead, copper, and zinc to the water.
The contribution of brass fixtures to first-draw lead and copper samples in tap water moni-
toring programs is now well accepted (Schock et al. 1996).
In studies, polyphosphates proved to be effective in suppression of corrosion of brass
and bronze as long as zinc was present at a concentration of 1 mg/L (Schock et al. 1996).
Orthophosphates also appear to control corrosion.
Copper. Copper (Figure 4-4) is currently the most common material for prem-
ise plumbing. Reasons include ease of installation, low cost, and corrosion resistance.
Although copper corrosion is rare, it can be severe, resulting in pinhole leaks and blue or
green water. Copper is much more resistant to uniform corrosion than the other metals
typically found in potable water systems except in soft waters with low pH. Also, hard
high-alkalinity groundwaters are particularly aggressive toward copper materials. Copper
is prone to erosion corrosion due to high velocity water, typically seen in recirculating hot
water systems where water flows past pipe bends. Impingement attack of copper by high
water velocities is one of the problems of copper pipe (Schock 1999).
Adjustment of pH has been shown to be very successful in controlling copper cor-
rosion and is usually the least expensive option. Anion concentrations (chloride, sulfate,
bicarbonate, orthophosphate) are purported to exert an influence on corrosion.
There are numerous articles on the effect of phosphate to control copper release in
real systems and from pipe-rig studies. Orthophosphate in sufficient dosage at the proper
pH can reduce copper corrosion. It has been suggested that 3 to 5 mg PO4 /L orthophosphate
may be necessary to achieve substantial improvements in cuprosolvency at approximately
pH 8.0, but perhaps only 1 to 3 mg PO4 /L at approximately pH 7.0 (Schock et al. 1995).
M58 book.indb 64 11/17/2010 4:24:52 PM

Figure 4-1 Asbestos–cement pipe
Figure 4-2 Bronze fitting
Figure 4-3 Brass faucet
Figure 4-4 Copper pipe
M58 book.indb 65 11/17/2010 4:25:00 PM

Figure 4-5 Galvanized pipe
Galvanized steel. Galvanized steel pipe (Figure 4-5) was the dominant material
used in premise plumbing and only replaced in the past few decades by copper and plas-
tic. In the United States and northern Europe, new installations of galvanized pipe are
rare. Galvanized pipe may be found in homes built prior to the 1980s (AwwaRF and TZW
1996). Galvanized steel pipe consists of a base steel layer with layers of iron/zinc alloy
that approach pure zinc at the interior surface of the pipe. Impurities in the iron/zinc alloy
include lead, barium, chromium, and cadmium. Galvanized pipe is expected to give good
service in hard or soft waters in the pH range of 7.0 to 8.5 but performs better in hard
waters.
The corrosion issues with galvanized pipe are complex. When the pipe is new, cor-
rosion products will be zinc and impurities of the galvanized layer such as lead and cad-
mium. The quality of the zinc coating plays a major role in corrosion. Problems with the
galvanization process can lead to pitting instead of uniform corrosion. Once the zinc lay-
ers have been corroded away, the pipe behaves as a black iron pipe. According to several
studies, orthophosphate effectively controls zinc solubility and galvanized steel corrosion
(Trussell and Wagner 1996). Silicates have also been shown to reduce galvanized pipe cor-
rosion, but most studies have focused on hot water systems. Adjustment of pH between 7.0
and 8.5 is also a very effective treatment.
Cast iron. Iron is the most common material in the distribution system and is usu-
ally found as cast iron, ductile iron, and steel. The forms of iron-based materials are dis-
tinguished by their nonferrous components. The major difference is that steel has lower
carbon and silicon content than either cast iron or ductile iron.
Unlined cast-iron mains (Figure 4-6) have been in use in the United States since
the early 1800s. Lined cast-iron mains have been installed since the 1970s. The linings,
intended to prevent or minimize corrosion, are usually asphaltic coatings over a cement
lining. Distribution systems older than 40 years may have a large portion of their distribu-
tion mains as unlined cast-iron. Corrosion of the iron-containing materials may be either
general or localized. The mechanisms of iron corrosion are covered in chapter 2.
Although chlorine has the greatest negative impact on new pipe, its impact dimin-
ishes as the pipe surface is passivated by corrosion products. Once corrosion products
have formed, chlorine can provide some protection from corrosion by reducing the for-
mation of ferrous iron, which is more soluble than ferric iron. Iron corrosion may also be
reduced through alkalinity and calcium adjustment or the addition of a corrosion inhibi-
tor (Clement et al. 2002).
Phosphate-based inhibitors are available as phosphoric acid, zinc or potassium
orthophosphate, zinc or sodium polyphosphates, and orthophosphate–polyphosphate
blends. Polyphosphates are known as cleaning agents and will remove corrosion prod-
ucts. They are also sequestering agents. They do not reduce iron corrosion but keep the
corrosion products in a less objectionable form. Orthophosphate, with or without zinc,
M58 book.indb 66 11/17/2010 4:25:01 PM

Figure 4-6 Unlined cast-iron main
Figure 4-7 Harvested lead service lines
Figure 4-8 Plastic water supply pipe
reduces corrosion by binding with the corrosion products and “hardens” the scale, keeping
it attached to the pipe wall.
Corrosion inhibitors containing silica have been shown to slow the oxidation of fer-
rous iron and reduce the hydrolysis of ferric iron. Thus, silicate chemicals may be used to
reduce iron corrosion.
M58 book.indb 67 11/17/2010 4:25:04 PM

Lead. In addition to lead pipes, sources of lead in household plumbing include pipe
jointing compounds, lead solder, and brass and bronze fixtures. The use of lead for ser-
vice lines was prohibited by the Lead Contamination Control Act of 1988. This law also
restricted the lead content of solder to less than 0.2 percent and less than 8 percent in
other materials in contact with drinking water. As of 1990, there were approximately 6.4
million lead connections and 3.3 million lead service lines in the United States (Schock
et al. 1996). Many water utilities began systematic replacement of lead service lines (Fig-
ure 4-7) with the promulgation of the LCR in 1991.
Corrosion control of lead-bearing materials may be accomplished through (1) the
physical removal of lead-containing materials and (2) water quality adjustment. The
removal of lead service lines is expensive. In addition, it is not always possible to remove
the entire lead service line or all lead-containing materials. The best option is usually to
approach lead corrosion on both fronts—remove as much lead-containing material as eco-
nomically and physically possible while adjusting water chemistry.
Zinc orthophosphate and orthophosphoric acid have been successfully used as cor-
rosion inhibitors for lead control in many systems. However, the effectiveness depends
on proper control of pH and DIC. Polyphosphates have been shown to be detrimental to
lead control because they are strong complexing agents for lead and calcium. Orthophos-
phate–polyphosphate blends have shown some success primarily due to the orthophos-
phate component. Silicates may also reduce lead solubility. New research has shown that
extremely high free chlorine residuals (>3.0 mg/L) may also prevent corrosion of lead, but
the chemistry is still not fully understood and may result in excessive DBPs. Stannous
chloride has received some attention as a possible corrosion inhibitor for lead. However,
stannous chloride is not currently a USEPA-approved method for corrosion control.
Plastic. Many service lines and some premise plumbing are being replaced with
plastic (Figure 4-8). Plastic lines may either be PVC, HDP, or chlorinated polyvinyl chlo-
ride. Plastic is a good choice of material for use in areas with aggressive water. A few local
codes restrict the use of plastic pipe. Plastic is resistant to corrosion but has some non-
corrosion-related disadvantages. Plastic pipe is less resistant to internal biofilm growth.
Buried plastic pipe can be difficult to locate as it does not conduct electrical current for
tracing. Plastic pipe is also more susceptible to permeation. Also, the issue of the potential
health effects of plasticizers has recently been raised.
A B
Source: National Research Council of the National Academies 2006.
Figure 4-9 Branched (A) and grid/loop (B) designs
M58 book.indb 68 11/17/2010 4:25:05 PM

Engineering Considerations
The design of pipes and structures in the distribution system can be as important as the
selection of materials. Some important design considerations include the following:
• Avoid dead ends and stagnant areas
• Provide adequate flushing of the system
• Select an appropriate flow velocity (5–7 ft/s [1.52 – 2.13 m/s] at maximum system
demand)
• Reduce mechanical stresses such as flexing of pipes and water hammer
• Avoid uneven heat distribution
• Reduce sharp turns and elbows
The two basic configurations for most distribution systems are the branch and the
grid/loop (Figure 4-9). In a branch system, smaller pipes branch off of larger pipes through-
out the service area such that water can only take one path from source to consumer. This
type of system is most commonly found in rural areas. In a grid/loop system, water may
take several pathways from source to consumer. This design can be found in urban areas.
Looping reduces some of the problems associated with stagnant water such as adverse
reactions with the pipe wall. Most systems will be a combination of the branched and
looped designs.
The purpose of a distribution system is to deliver water to all customers of the system
in sufficient quantity for potable drinking water and fire protection. Although only a small
percentage of water is used for fire-fighting purposes, the sizing of water mains is partially
based on fire protection requirements to ensure sufficient flow for a minimum period of
time. Generally, 75 percent of the capacity of a typical drinking water system is devoted
to fire fighting (NAS 2006). The effect is that there may be long residence times before the
water reaches the consumer, increasing the available time for reactions between the pipe
wall and the water in the pipe as well as creating an environment to encourage the growth
of microorganisms. In the absence of smaller distribution systems, water utilities have to
implement flushing programs and to increase dosages of disinfectants to maintain a resid-
ual at the ends of the distribution system. One suggested alternative is to have dual dis-
tribution systems—one dedicated to potable water use and the other to fire fighting. The
Irvine Ranch Water District in Irvine, Calif., is one of the most widely known utilities using
a dual distribution system. Dual systems are most advantageous in new communities.
Erosion corrosion has been observed in pipes with high flow velocities or where
abrupt changes in flow directions exist. The abrasive action of the fluid can scour the pipe
scale and remove any protective covering that may have formed on the pipe wall. High
flow velocities may increase the rate at which an oxidant species comes in contact with
pipe surfaces, thus increasing corrosion. Water distribution systems designed to operate
at lower flow rates will have reduced turbulence and, therefore, reduced erosion of the
protective layer.
It has been previously mentioned that compatible materials should be used to reduce
galvanic corrosion. However, the use of compatible materials is not always possible. Dielec-
tric insulators can be used to isolate dissimilar metals. The purpose of a dielectric insula-
tor is to provide a nonconductive barrier between two conductive (metallic) components
of a pipeline or piping system. The corrosion control philosophy and long-term integrity of
many pipelines and piping systems depend on maintaining full metallic isolation at one or
more locations over the life of the equipment. Insulator sites need to be selected early in
the project design so that provisions for their installation, testing, and maintenance access
may be properly incorporated.
M58 book.indb 69 11/17/2010 4:25:05 PM

The galvanic couples that commonly occur in the water and wastewater industries
can be rated according to severity. Examples of dangerous galvanic couples in the water
industry are listed in Table 4-3.
The most common insulators used in the water industry are
• Insulating flange kits
• Casing insulators (centralizers)
• Dielectric unions
• Insulating bushings
• Structural design (pipe penetrations through concrete)
• Nonmetallic or link-seal–type sleeves
• Nonmetallic pipe spools
• Dielectric tubing connectors
• Direct current decoupling devices
• Insulating couplings
• Monolithic insulators
Each insulator has advantages and disadvantages depending on its application. Some
municipal codes regulate the use of dielectric insulators in water systems.
System Maintenance
Routine system maintenance can reduce corrosion and/or remove the by-products of cor-
rosion. Flushing is performed by isolating sections of the distribution system and open-
ing up fire hydrants or flushing valves to cause a large volume of flow to move through
isolated pipes so that a scouring action is created, thus removing any material buildup
from the pipe. The flushing should be of sufficient velocity to suspend loose sediment.
Unidirectional flushing procedures are preferred over conventional flushing because con-
ventional flushing may not sufficiently remove corrosion by-products. Flushing eliminates
the buildup of iron and manganese scales that can bind and accumulate other metals in
the pipeline. Flushing is also a control technique for MIC where biological deposits are
removed. Combining flushing with the use of higher disinfectant residuals or alternative
disinfectants can lead to the removal of microorganisms involved in MIC.
Cleaning and lining of old pipe can be an effective way to reduce metals release and
rehabilitate old pipe. Cleaning procedures include the use of proprietary acids and surfac-
tants or mechanical scrapers to remove scale and deposit.
Table 4-3 Galvanic couples in the water industry that are dangerous
Anodic Site Cathodic Site
Dielectric-coated ductile-iron pipe Mortar-coated steel pipe
Steel pipe (dielectric coated) Copper services, blowoffs
Steel pipe (mortar coated) Copper services, blowoffs
Polyethylene-encased ductile-iron pipe Mortar-coated steel pipe
Source: Based on DeCarlo 2004.
M58 book.indb 70 11/17/2010 4:25:05 PM

Figure 4-10 Application of cement–mortar lining
Figure 4-11 Example of pipe before and after cleaning and lining
Pipe Coatings and Linings

In addition to selecting the proper piping material to minimize corrosion, one technique to
keep corrosive water away from the pipe wall is to line the wall with a protective coating.
With new pipe, these linings are usually applied during the pipe manufacturing process
or in the field when the pipe is installed. Some linings can be applied even after the pipe
is in service.
The most common pipe linings are coal-tar enamels, epoxy paint, cement–mortar,
and polyethylene. It is common practice to reline cleaned pipe to protect the new exposed
pipeline material. The most common technique is to apply cement–mortar. Figure 4-10
illustrates one process to apply a cement–mortar lining. Spray-on epoxy lining is espe-
cially useful in water that is low in hardness, which can cause a cement lining to deterio-
rate. Figure 4-11 illustrates iron pipe that has been cleaned and relined.
Cement–mortar has an expected life of 50 years and costs $1 to $3 per inch diameter
per foot (2007 data). Epoxy lining is estimated to last more than 75 years but costs $9 to
$15 per inch diameter per foot (2007 data). Costs will vary by geographical location as
some materials are more readily available in urbanized areas.
Figure 4-12 illustrates steel pipe with an epoxy lining. Table 4-4 summarizes the most
commonly used pipe coatings and their advantages and disadvantages.
M58 book.indb 71 11/17/2010 4:25:06 PM

Source: USACERL Technical Report 99/39.
Figure 4-12 Steel pipe with epoxy coating
Table 4-4 Pipe wall linings

Material Use Advantages Disadvantages
Hot-applied coal • Lining for steel pipes • Long service life (>50 years). • Need to reapply to welded
tar enamel • Good erosion resistance—to areas.
silt or sand. • Extreme heat may cause
• Resistant to biological cracking.
attachment. • Extreme cold may cause
brittleness.
• May cause an increase in
trace organics in water.
Epoxy • Lining for steel or • Smooth surface results in • Relatively expensive.
ductile pipes, applied reduced pumping costs. • Less resistance to abrasion
to some lead service • Formulated from components than coal tar enamel.
lines in the field approved by the Food and • Service life <15 years
Drug Administration.
Cement–mortar • Standard lining for • Relatively inexpensive. • Rigidity of lining may lead to
ductile-iron pipes, • Easy to apply. cracking or sloughing.
sometimes used in • Calcium hydroxide release • Thickness of coating reduces
steel or cast-iron pipes may protect uncoated metal cross-sectional area of
in pipe joints. pipe and reduces carrying
capacity.
Polyethylene • Lining used in ductile • Long service life (50 years). • Relatively expensive.
and steel pipes • Good erosion resistance to • Emerging health concerns
abrasives. related to plasticizers
• Good resistance to bacterial potentially leaching into the
corrosion. water.
• Smooth surface results in
reduced pumping costs.
Source: USEPA 1984.
M58 book.indb 72 11/17/2010 4:25:06 PM

Table 4-5 Common corrosion control chemicals

Bulk
Chemical Name, Formula, Commercial Density, Solubility,
and Common Name Common Forms Strength lb/ft3 g.100 g H2O
Calcium carbonate White or gray; granular or Approx. 95% 100–115 0.0014 at 25°C
CaCO3 powder CaCO3
High calcium limestone
Calcium hydroxide White powder 72–74% as CaO 25–35 0.19 at 0°C
Ca(OH)2 0.15 at 30°C
High calcium hydrated lime, 0.08 at 100°C
slaked lime
Calcium oxide White; lump, pebble, granular, 93–98% CaO 55–60 Converts to
CaO or pulverized Ca(OH)2 in
High calcium quicklime, solution
burnt lime, unslaked lime
Dolomite hydrated lime White powder 46–48% CaO 25–35 0.18°C
Ca(OH)2 MgO 33–34% MgO
Dolomite limestone White, gray, or tan; granules Approx. 52.6% 105–120 0.002 at 25°C
CaCO3 MgCO3 or powder CaO3
42.5 MgCO3
Dolomite quicklime White, light gray, or tan; lump, 55–57.5% CaO 55–60 Converts to
CaO MgO pebble, or pulverized 37.6–40.8% MgO Ca(OH)2 in
solution
Sodium bicarbonate White powder 99% 135 6.9 at 0°C
NaHCO3 18.4 at 100°C
Bicarb, baking soda
Sodium carbonate Anhydrous; white powder >98% 35–65 7.1 at 0°C
Na 2CO3 18 at 20°C
Soda ash 27.6 at 30°C
45.5 at 100°C
Sodium hydroxide Anhydrous; white solid beads, 98.9% (NaOH) 133 42 at 0°C
NaOH flakes, or liquid 76% (Na 2O) 347 at 100°C
Caustic soda 50%
Sodium silicate Opaque, viscous liquid 38–42% Be Complete
Na 2O(SiO2)
Water glass
Source: AwwaRF and TZW 1996.
In 1999, the US Army Corps of Engineers published the results of a 12-week study
comparing the application of in-situ epoxy coating to chemical treatment for the control
of lead and copper corrosion (Hock et al. 1999). The results indicated that the epoxy coat-
ing was as effective as chemical treatment in controlling lead corrosion and that epoxy
coating reduced copper corrosion better than the application of corrosion inhibitors. The
results were duplicated in a full-scale study in the water distribution system of a small
army base. The application of the epoxy may be performed in the field to lead service
lines and may be an economical alternative to lead service line replacement. Water quality
testing of the in-situ epoxy–lined pipes demonstrated very little leaching of organic con-
taminates as long as the pipes were allowed to cure completely and the lines were flushed
prior to first use.
M58 book.indb 73 11/17/2010 4:25:06 PM

CHEMICAL TREATMENT______________________________________
Corrosion control can be characterized by two general approaches: (1) forming a precipi-
tate in the potable water supply that deposits onto the pipe wall to create a protective coat-
ing, and (2) causing the distribution system pipe material and the water supply to interact
and form metal complexes on the pipe surface (USEPA 1984). The latter mechanism, often
called passivation, is where pipe material and existing scale deposits are used to form a
barrier film of less soluble metal carbonates or phosphate compounds on the inner pipe
surface. This barrier layer of metal complex isolates the plumbing materials from the
water supply and minimizes dissolution.
In many cases, the easiest way to reduce corrosion is to modify the water quality at
the treatment plant or well house. Chemical treatment options include pH adjustment, DIC
adjustment, and application of corrosion inhibitors. The USEPA Revised Guidance Man-
ual for Selecting Lead and Copper Control Strategies (2003) includes several sets of flow
charts to help select viable chemical treatment options based on water quality characteris-
tics. More than one chemically viable treatment option may be available to a water system.
The tools presented in chapter 3 will help select the best chemical treatment option.
Different chemicals have different water quality and corrosion control attributes.
Each treatment also has various operational aspects that must be considered when evalu-
ating different corrosion control options. Tables 4-6 and 4-7 summarize water quality and
operational aspects for various treatment chemicals.
The type of chemical used for corrosion control treatment will determine what type
of chemical feed system is used. A typical liquid chemical feed system is illustrated in
Figure 4-13, and a dry chemical feed system is illustrated in Figure 4-14. A basic chemi-
cal injection system should perform three functions: chemical storage, chemical injec-
tion, and adequate system controls to consistently provide the appropriate chemical dose.
Chemical storage includes a bulk storage tank and sometimes a smaller tank referred to
as a day tank. The bulk storage tank or tanks should be sized based on (1) the desired
supply of chemical, (2) amount of physical space, and (3) shipment size. Preferences vary,
but some utilities size bulk storage tanks for several weeks’ worth of storage while other
utilities prefer a supply of 3 or more months. Ten States Standards recommends 30 days at
average dose and average flow. Some states require a 30-day supply at maximum storage.
Some chemicals are less expensive when ordered in large quantities.
Figure 4-13 Typical liquid chemical feed system
M58 book.indb 74 11/17/2010 4:25:07 PM

Figure 4-14 Typical dry chemical feed system
Table 4-6 Treatment chemical water quality and corrosion control aspects
Water Quality and Corrosion Water Quality and Corrosion
Chemical Control Advantages Control Disadvantages
Soda ash (Na 2CO3) • Increases total alkalinity of 0.94 mg/L • 0.43 mg/L sodium added per 1 mg/L soda ash.
CaCO3/L for 1 mg/L Na 2CO3, increases More than 10 mg/L Na added and total 20 mg/L
DIC and pH to above 8.3. Na should be avoided.
Sodium bicarbonate, • Increases both pH and alkalinity but • Cannot raise pH > 8.3.
baking soda (NaHCO3) not as much as the same dose of soda
ash. Alkalinity increases by 0.59 mg/L
CaCO3/L for 1 mg/L dose.
Hydrated lime Ca(OH)2 • Increases pH and calcium for soft water. • Overdose of lime in soft waters with low free
10% lime milk or 1% lime CO2 will increase pH above the acceptable
water range. Therefore, addition of lime to soft waters
is sometimes accompanied by CO2.
• Quality of lime varies. Lime can contain
contaminants, such as heavy metals and high
concentration of aluminum. If high dose of lime
is needed, the increase in Al concentration can
be as high as 0.5 mg/L. The contribution of Al
from lime can be minimized by good turbidity
control.
• Often increases turbidity of treated water.
Quicklime • Same for hydrated lime. • Same for hydrated lime.
CaO
Table continued on next page.
M58 book.indb 75 11/17/2010 4:25:08 PM

Table 4-6 Treatment chemical water quality and corrosion control aspects (continued)
Caustic soda • Adds hydroxide alkalinity. • Even small overdose of caustic soda in very
NaOH soft waters with low CO2 increases pH to
excessively high values. For soft low-alkalinity
water, NaOH is recommended only when used
in conjunction with CO2 or when CO2 is present
in the water.
• Adds sodium to the water.
• No carbonate alkalinity increase unless free CO2
is present in the water or added.
Lime and CO2 • CO2 should be added before lime.
Lime and soda ash • Used in Seattle to increase pH from
around 6.0 to 8.0, with an increase in
alkalinity from 5 to 17 mg/L CaCO3.
Reduced corrosion of old Cu surfaces
by 70%, as well as reduced the release
of Cu, Pb, Zn, and Fe in standing water
samples from homes.
Limestone filter • Used in smaller systems to increase the • CO2 gas is frequently used to increase the rate
(CaCO3, hardness, pH, and alkalinity of soft of CaCO3 dissolution and/or to stabilize the pH
CaCO3 –MgO + CO2) water. from CaCO3 –MgO media filters. Addition of
High Ca 99% CaCO3 CO2 is often necessary to prevent excessive pH
Dolomite 54% values in low-alkalinity water treated with half-
CaCO3 46% burned dolomite filters CaCO3 –MgO.
MgCO3 • Limestone may vary in quality and may contain
a number of impurities: clay, silt and sand,
iron, phosphorus, sulfur, and trace amounts
of manganese, copper, sodium, and arsenic
(USEPA 1986).
Orthophosphate • Effective at moderate pH 7.0–8.0 for • Insufficient dose may accelerate corrosion
products copper control (Reiber et al. 1997). or may not avoid pitting or other localized
Orthophosphate can reduce lead corrosion. Very pH-dependent with copper
solubility by building up a lead corrosion protection lost at slightly acidic pH
phosphate layer that has a low solubility. values. Orthophosphate dosages of 1–5 mg
Reduced copper corrosion rate in cold P/L had only minimal effects in reducing zinc
waters. Effectively reduces the general corrosion in Seattle tap water at pH 8.0.
corrosion of the zinc layer of galvanized • Orthophosphate can stimulate algal growth in
steel. Can reduce corrosion rate of steel waters receiving wastewater discharges.
and impact iron-release rate. Small • Ineffective for copper control outside pH range
doses of 0.5–1.0 mg/L reduced iron of 7.0–8.0 (Reiber et al. 1997).
release by 67%.
• Normally have an inhibiting effect if the
corrosion products or the corroding
metal can form insoluble phosphates
and improve the protective scale by
increasing its impermeability and
adherence.
M58 book.indb 76 11/17/2010 4:25:08 PM

Polyphosphate products • Prevent red water for iron corrosion. • Unknown how and why they work. May
• Maximum benefit for lead control will increase zinc, copper, and lead corrosion.
occur in the lowest alkalinity waters Polyphosphates may increase lead levels by
(Reiber et al. 1997). complexation. Usually have little or no effect on
either corrosion inhibition or the reduction of
metal release.
• Adverse effects on lead corrosion control for
waters with alkalinity between 30 and 74 mg/L
CaCO3. Variable results for copper control
above pH 7.8 (Reiber et al. 1997).
• Polyphosphates are capable of solubilizing metal
oxides and dissolving some of corrosion scales
and likely not suitable for lead and copper
control (Reiber et al. 1997).
Zinc orthophosphate • Historically used for corrosion control of • Many utilities use a combination of zinc sulfate
iron surfaces. and phosphoric acid for LCR compliance. The
mechanism of zinc action on lead and copper
surfaces has never been quantified. Zinc
orthophosphate has not offered significant
benefit versus simple orthophosphates. Zinc
is currently not allowed in corrosion inhibitor
products in much of Europe and is becoming
increasingly restricted in North America. Zinc
in concentrations as low as 0.01 mg/L has been
toxic to trout and certain other fish.
Phosphate blends • Dose dependant on pH and Ca • Insufficient doses can aggravate pitting
concentration. Low doses are used corrosion. Blends with small proportion of
to avoid soft scale formation and orthophosphates have been shown to increase
discolored water. corrosion rates and iron release.
• Phosphates form protective films on the • Corrosion control may require 3–6 months of
bare metal surface by chemical reaction dosing.
with either the corrosion products or • Variable corrosion inhibition has been
the bare metal. found for mixtures of orthophosphates and
polyphosphates.
• Proprietary phosphate blends do not outperform
less expensive generic orthophosphate.
• Polyphosphate inhibitors had adverse effects
on lead corrosion above pH 7.0 likely due to
complexation and release of lead corrosion by-
products (Reiber et al. 1997).
M58 book.indb 77 11/17/2010 4:25:08 PM

Silicates • Poorly soluble in cold water. Soft waters • Corrosion must occur before the metal surface
(N2O:SiO2 mixtures) with low pH are most likely to be treated can be protected due to mechanism for
with silicates at 4 to 30 mg/L. A major inhibition being the film of ferric or another
benefit of using silicates appears to be metal oxide and silicate.
the accompanying pH increase. It is • The protective films will gradually break down
often more economical to increase the and protection will cease within a short time if
pH to between 7.5 and 8.0 by adding the dosage is stopped.
caustic soda or soda ash to lower the • Silicates do not inhibit zinc corrosion beyond the
silicate requirement. The corrosion effect of increased pH.
protection appears to be caused mainly • Silicates can stimulate certain biologic growth in
by the pH increase and film formation. low-mineralized waters.
May reduce red-water problems, more so • Silicate component of caustic silicates (apart
than pH or pH and alkalinity adjustment from pH adjustment) had little evidence of any
alone. inhibitory impact on either lead or copper.
• Protection by the silicates occurs by • Limited evidence suggests that silicate addition
forming a thin layer over the corroded without pH control may exacerbate copper
metal layer. The films do not build up corrosion (Reiber et al. 1997).
and are self-limiting.
• Generally less effective than
orthophosphates. Often observed
effects depend on the pH increase
caused by the alkaline character of the
silicate compound.
Phosphate–silica • Provide corrosion protection as good • Even in the case of very high pH of 8.7, when the
mixtures as phosphate alone but at reduced iron release rate decreases to very low values,
phosphate concentrations. Dual the corrosion rate remained relatively high in
mechanism may provide better the low-mineralized water. The result of this
corrosion protection. high corrosion rate and low iron-release rate
• Phosphate–silicate blends often show is that a large quantity of ferric precipitates is
good corrosion reduction in steel pipe. deposited in the pipes.
M58 book.indb 78 11/17/2010 4:25:08 PM

Table 4-7 Chemical operational aspects

Chemical Handling and Use Advantages Handling and Use Disadvantages
Soda ash (Na 2CO3) • Relatively easy to work with. Mainly • Hydrolyzes by absorbing water from air, which can make
used in powdered form, delivered it cake. Sacks should not be stored more than 6 months.
in bulk or in sacks. The chemical • The chemical cost is higher than that of lime (roughly 10
readily dissolves, is considered times), and pH control is needed.
nonhazardous and relatively easy to
handle and to dose; it is often used
in smaller water facilities.
• Solubility is 71 g/L at 0°C and
increases with temperature.
Sodium • Nonhazardous chemical; easy to work • The chemical cost is higher than others, about 40 times
bicarbonate, with. higher than CaCO3. Needs pH control.
baking soda • Solubility is 69 g/L at 0°C and
(NaHCO3) increases with temperature.
Hydrated lime • Relatively inexpensive chemical. • May clog dosing equipment.
Ca(OH)2 • Solubility is 1.9 g/L at 0°C and • Less hygroscopic and much easier to store than
10% lime milk or decreases slightly with temperature. quicklime, but carbonation can cause deterioration.
1% limewater • Low winter temperatures in an unheated building could
lead to freezing or caking.
• Requires large investment in chemical dosing equipment.
• Need to have sedimentation tank or saturator to prevent
lime particles from going into the treated water. In
other cases, deep bed filtration is used after lime
addition.
• May cause dust problems if special precautions are not
taken.
• Turbidity and pH control are necessary; pH control is
difficult for low-alkalinity water.
Quicklime • Solubility same as for hydrated lime. • Requires lime slaker to produce hydrated lime. Usually
CaO used in larger water facilities. Lime is difficult to use
because of its low solubility and high hygroscopy, which
increase the risks of caking and blockages.
• Must be stored in moisture-proof areas and used within a
few weeks of manufacture to avoid absorbing water and
CO2 and forming calcium carbonate cake.
Caustic soda • Small capital and investment cost. • Hazardous chemical. Severe burns can occur to skin
NaOH Easy to dose. or eyes on exposure. Safe handling of tanks, pumps,
• Solubility is 420 g/L at 0°C and piping, and dosage equipment must be implemented as
increases with temperature. specified by building and fire codes. Strict control of the
dosage is necessary.
• Low winter temperatures in an unheated building could
lead to freezing or caking. The freezing point of 30%
solution is still 10°C.
Lime and CO2 • Used in many conventional treatment
plants—e.g., Tolt plant in Seattle
and plants in Finland, Sweden, and
Norway—having soft low-alkalinity
water. The same treatment is also
being used in desalination plants in
Saudi Arabia.
• Selecting the right sequence and
proportion of chemicals can control
the desired concentrations of
bicarbonate, pH, and calcium.
M58 book.indb 79 11/17/2010 4:25:08 PM

Table 4-7 Chemical operational aspects (continued)

Lime and soda ash • Used to be applied by Seattle
Limestone filter • Eliminates the need to feed chemicals • No control over pH, which depends on the flow rate and
(CaCO3, CaCO3 – and cannot result in overdosing. could vary more than 1 unit. The rate of dissolution of
MgO + CO2) Claimed to be low cost and simple CaCO3 may vary depending on impurities in limestone
High Ca 99% operation. (USEPA 1986).
CaCO3 • Solubility of high calcium limestone • Equilibration of the column with atmospheric CO2 can
Dolomite 54% is 14 mg CaCO3/L at 25°C, and have a significant impact on pH (USEPA 1986).
CaCO3 46% MgCO3 half-burned dolomite is 1.8 g/L. • Fairly large investment in an alkaline media filter or a
Mackintosh et al. (2002) report limestone contactor. The dissolution rate for limestone
the following design guidelines: is low. Half-burned dolomite has a much higher
(1) cylindrical configuration with dissolution rate and is, therefore, preferred in many
a height to diameter ratio at least cases. Requires a better pH control system. Very high
1:1; (2) minimum bed depth 2 m pH values can occur.
and minimum clear water depth • Needs to be backwashed when clogging reduces the flow
of 0.2 m; (3) completely enclosed capacity.
structure with access hatch on top • Norwegian experience indicates that pH of 3.5 to 4.0 is
for limestone addition; and (4) use of required to ensure sufficient dissolution of CaCO3 by
false-bottom feed system. adding CO2 or very high doses of coagulant prior to the
limestone filter (Odegaard 1999). Grain size, pH, and
contact time influence the dissolution of the marble.
The empty-bed contact time in the marble part of the
filter must be at least 15 to 25 min. Limestone filters
in Norway are equipped with backwash (relatively
high backwash water consumption is needed, 6–11%)
(Odegaard 1999).
• Because no standard installations exist in California,
limestone filters should be considered an experimental
process. Mackintosh et al. (2002) reported from South
African experience that poor design or poor operation
have resulted in some instances in noneffective
limestone contactors.
• Disadvantages presented by Mackintosh et al. (2002):
(1) at least two contactor units per installation;
(2) limestone contact time at least 20 min; (3) maximum
loading rate 10 m/h; (4) need for piezometers to measure
pressure loss across the limestone bed to indicate when
“down flushing” of insoluble fines is required; (5) piping
to allow overflow, flushing to waste in both upflow and
downflow mode; (6) location after clarification, filtration,
and disinfection (turbidity <1 ntu, iron <0.1 mg/L,
aluminum <0.15 mg/L), otherwise excessive fouling and
failure of the process may occur; (7) only selected and
tested limestones are acceptable for use; (8) ensure
equal flow distribution; (9) need to flush the fines after
each addition of limestone to remove milky yellow
turbid water; (10) down flushing of fines to waste at least
monthly; (11) pre- and poststabilization levels need to
be monitored; (12) when adding limestone water must
be running at maximum allowable rate or higher and
wasted; and (13) limestone must be flooded at all times.
Silicates • Sodium silicate has been used • The cost is higher than that of phosphates and bimetallic
(N2O:SiO2 in point-of-use applications at inhibitors.
mixtures) individual buildings and for
industrial processes, and in
municipal supplies in the eastern
United States and Canada.
• Readily soluble.
M58 book.indb 80 11/17/2010 4:25:08 PM

Table 4-7 Chemical operational aspects (continued)

Various inhibitors • Orthophosphates and polyphosphates • Pilot-plant studies are often required because the level
are readily soluble. of corrosion reduction can vary substantially with
• The inhibitor chemical costs are inhibitor.
relatively minimal. • Minimum concentration necessary for long-term
operation has to be determined. Stopping inhibitor
dosage revives the corrosion problems. If good corrosion
protection was provided, a certain time without dosing
can be accepted without risk. It is necessary to control a
reduction in inhibitor dosage by a monitoring program.
The treatment chemical is injected into the water at an appropriate dosage rate.
There are three basic methods for controlling the injection rate: flow pacing, pH or con-
centration pacing, and operation actuated by the well pump motor startup in groundwater
systems. The chemical is transferred from the tank using either flexible tubing or rigid
piping. The type of piping used will be determined by the chemical. Strong alkaline liquids
such as sodium or potassium hydroxide are notorious for leaking at connections and fit-
tings. Rigid stainless steel piping with compression fittings have been successfully used
for these chemicals at many installations. Flexible tubing is less expensive than piping,
but it is difficult to eliminate leaking at the tubing connections in flexible tubing.
The chemical feed system must contain controls. Basic controls include high-level
and low-level sensors located in the storage tanks. The high-level sensor can trigger an
alarm that can help the operator avoid overfilling the tank; and the low-level sensor can
trigger an alarm that the tank needs to be filled. The system should also have a chemical
containment area capable of containing 110 percent of stored chemical volume.
Adjustment of pH and/or DIC Adjustment

The adjustment of pH is the most common method for reducing corrosion in distribution
systems. The water pH is the major factor that determines the solubility of most pipe
materials and the films that form from corrosion by-products. Most materials in the distri-
bution system dissolve faster at lower pH.
The relationship of pH, DIC, alkalinity, carbon dioxide, and ionic strength governs
the solubility of metal carbonates, which are commonly involved in protective scale for-
mation on interior pipe surfaces. The type of alkalinity present and the pH of the water are
dependent on each other.
The solubility of metals typically decreases as pH increases. Increased pH may, how-
ever, prove troublesome for water that contains elevated levels of iron and manganese,
increased THM concentrations, and excessive calcium carbonate precipitation; these will
cause hydraulic issues and increased pumping costs. Iron and manganese are most likely
to precipitate out and cause aesthetic issues. Figure 4-15 illustrates the reduced solubil-
ity of manganese and iron as pH increases. The soluble forms of iron and manganese are
represented by Fe+2 and Mn+2 on the diagram. A phosphate-sequestering agent is usually
used in these situations.
Caustic soda (sodium hydroxide), potassium hydroxide, soda ash, limestone contac-
tors (calcite filters), and aeration/air stripping (where sufficient dissolved CO2 exists in the
water) are the principal methods for increasing the pH. Soda ash, potash, and limestone
contactors also increase DIC; aeration decreases DIC in the process of increasing pH.
Sodium/potassium hydroxide. When sodium hydroxide or potassium hydroxide is
used, dosing must be strictly controlled. Caustic is recommended only for waters with DIC
greater than 5 mg/L. In very soft waters with low carbon dioxide, even a small overdose
M58 book.indb 81 11/17/2010 4:25:09 PM

of caustic will increase pH to very high values. For soft low-alkalinity waters, caustic
is recommended only when used in conjunction with carbon dioxide. Caustic is usually
delivered in a 50 percent solution but may be delivered or diluted to lower concentrations.
Caustic is a hazardous chemical, and severe burns can occur on exposure. Caustic has a
low freezing temperature. A 50 percent solution has a freezing point of 12°C, and a 30 per-
cent solution has a freezing point of 10°C. Caustic feed systems at a minimum should
include an eye washing system, full shower, eye goggles, protective gloves, boots, aprons,
easy-to-handle barrels, and chemical containment areas.
Soda ash/potash. Soda ash, or sodium carbonate, and potassium carbonate (“pot-
ash”) are dry compounds that are relatively safe to handle compared to caustic. These
carbonate chemicals will not cause skin irritation. These chemicals are recommended for
raw waters with DIC levels between 2 and 25 mg/L. When soda ash or potassium carbonate
is added to a water, there is an increase in DIC as well as in pH—approximately 0.94 mg/L
per 1 mg/L dose. Because soda ash and potassium carbonate are safe to handle, they are
strongly recommended as the pH adjustment chemical for schools, condominiums, or
any facility where technical resources are limited. They dissolve more easily than lime.
Potassium carbonate is more expensive than soda ash but is more soluble and easier to
handle. Therefore, many very small water systems have found potassium carbonate easier
to implement than soda ash.
Sodium bicarbonate. Sodium bicarbonate is a nonhazardous chemical with which
it is easy to work. This chemical is applied to waters with DIC less than 5 mg/L. When
sodium bicarbonate is added to a water, there is an increase in DIC as well as pH—approx-
imately 0.59 mg/L per 1 mg/L dose. This chemical is limited, however, in that it cannot
increase pH above 8.3. Some utilities use both soda ash or caustic and sodium bicarbonate
together if a significant increase in pH and alkalinity is needed.
Lime. Lime is available as either hydrated lime (CaOH) or quicklime (CaO). The use
of quicklime requires a lime slaker to produce hydrated lime. Quicklime is usually used in
larger facilities because of lower chemical costs. It must be noted that lime can be difficult
to use because of its low solubility and high hygroscopy (i.e., attraction to moisture in the
atmosphere). Caution should be taken to minimize moisture in quicklime storage areas and
silos. Most quicklime is used within a few weeks of chemical manufacture as it may form
calcium carbonate cake by absorbing moisture and carbon dioxide. Dry hydrated lime
is less hygroscopic and is easier to store than quicklime. Specifications usually require
quicklime and hydrated lime to contain 90 percent and 68 percent available calcium oxide,
respectively.
Fe(OH)+2
0.8 H2O=½O2+2H++2e− 0.8
MnO+4
MnO2
0.4 0.4
Fe(OH)3 Mn
OO
H
E h (volt)
0 Fe +2
0 Mn+2
Mn3O4
FeCO3 MnCO3
-0.4 -0.4
Mn(OH)2
Mn(OH)–3
Fe(OH)2
Fe
-0.8 -0.8
2 4 6 8 10 12 14 2 4 6 8 10 12 14
pH pH
Adapted from USEPA
Figure 4-15 Iron and manganese solubility
M58 book.indb 82 11/17/2010 4:25:09 PM

Lime may contain contaminants such as heavy metals including aluminum. Depend-
ing on the quality of lime and the amount used, it is possible to increase the aluminum level
by as much as 0.5 mg/L. Lime may also lead to excessive finished water turbidity if the lime
does not dissolve completely. Some utilities will use lime for gross pH adjustment and then
fine tune the pH with caustic.
A limestone contactor is usually an enclosed filter containing crushed high-purity
limestone. This treatment is usually used by small systems. As the water passes through
the limestone, the limestone dissolves, raising the pH, calcium, alkalinity, and DIC of
the water. The system is very simple and requires very little maintenance as it does not
require any pumps or continuous addition of limestone. Occasionally the limestone must
be replaced. When obtaining a design for a limestone contactor, it is important to ensure
that the contactor is adequately sized to produce sufficiently high pHs for the range of
flow rates and temperatures encountered during plant operation. The most successful are
operated in an upflow rather than a downflow mode. To use this treatment, a water system
should have pH less than 7.2, calcium less than 60 mg/L, and alkalinity less than 100 mg/L.
Iron and manganese may interfere with treatment because they can coat the limestone
interfering with dissolution.
Most water utilities will use a lime slaker. There are basically three types of lime
slakers available on the market.
• Slurry detention slakers
• Paste slakers
• Ball mill slakers
Figure 4-16 Slurry slaker
M58 book.indb 83 11/17/2010 4:25:09 PM

Lime Feed
Cut-off Spray
Header
Gearbox
Weir M
Slaking
Motor
Discharge
Chamber Vent
Overflow
Slaking Chamber Pipe
Slaked Lime (Slurry)

Discharge
Figure 4-17 Paste slaker
Vent Feed
Discharge
Agitated
Media
Level
Recirculation
Drain
Figure 4-18 Ball mill lime slaker
A slaker must mix the correct amount of quicklime (CaO) and water, must hydrate
the quicklime, and must separate the impurities and grit from resultant calcium hydrox-
ide slurry.
A slurry slaker generally uses an initial lime-to-water ratio ranging from 1 to 3.3 to
a ratio of 1 to 5 depending on the make of equipment and quality of CaO and water. Typi-
cally a slurry slaker, sometimes called a detention-type slaker, is comprised of two cham-
bers. The first chamber is called the slaking chamber, where lime and water are mixed.
The second chamber is usually used as a grit removal chamber. The lime slurry flows by
gravity from the first chamber to the grit chamber. The slurry viscosity is reduced in the
second chamber by the addition of cold water to allow the heavier grit to settle to the bot-
tom of the second chamber where the grit is elevated and discharged by a screw. Figure
4-16 shows one such slaker.
M58 book.indb 84 11/17/2010 4:25:10 PM

The slurry slakers are generally designed for a retention time of 10 min at full rated
capacity. This retention time means that from the time a particle of CaO enters the slaker
until it exits into the grit remover takes an average of 10 min.
Slurry slakers are available in a variety of sizes ranging from 150 lb/hr to 15 tons/hr.
Slurry slakers are also available with external vibratory grit separation screens. Because
the technology for slurry or detention-type slakers has existed longer, those slakers are
the most common types in Europe and the United States. Paste, and more recently batch,
slakers are becoming more common.
Paste slakers, as the name implies, slake the lime in a paste form. The lime-to-water
ratio is generally 1 to 2.5. Paste slakers are compact in size and are designed for a reten-
tion time of 5 min in the slaking chamber. In a paste slaker, because the hydroxide paste
is too heavy to flow by gravity, a pair of horizontal rotating paddles pushes the paste for-
ward toward the discharge point. Once the paste exits the slaking chamber, it is diluted to
approximately 1 part lime to 5 parts water. This dilution allows grit separation by gravity
or by an external vibrating grit screen. The slurry consistencies from a paste slaker and a
slurry slaker are exactly the same after dilution for grit removal.
Paste slakers are available in sizes ranging from 1,000 lb/hr to 8,000 lb/hr. Paste slak-
ers are mostly used in the United States. Figure 4-17 depicts a paste slaker.
Ball mill slakers are an adaptation of ball mills, which originally were designed for
wet and dry grinding, to the job of slaking. Two types of ball mills are used for slaking,
horizontal and vertical. Ball mill slakers are generally used where
• Capacity required is too large for other types of slakers
• Due to zero discharge conditions at the site, no grit discharge is allowed
• When water available is too high in sulfates or sulfites for regular slakers
The ball mill slakers are much more expensive than paste or slurry slakers. They are
available in sizes ranging from 1,000 lb/hr to 50 tons/hr. Figure 4-18 shows an attritor-type
vertical ball mill lime slaker.
The ball mill slakers are equipped with an external classifier, which separates
slurry from the oversized grit and impurities. The oversize grit is recycled back into the
mill for regrind.
Aeration. To use aeration for corrosion control treatment, a water system should have
a groundwater source or stratified surface water source. Aeration systems can increase
the pH of groundwater systems or stratified surface water systems by removing oversat-
urated carbon dioxide. A stratified surface water has different density and water quality
layers as a result of temperature changes over the summer season. The bottom layers often
lose dissolved oxygen and have elevated levels of carbon dioxide, iron, and manganese
because the stratification prevents diffusion of dissolved oxygen from the upper water to
the lower. Aeration is the only pH adjustment method that does not add a chemical to the
water and the only one that can reduce excess DIC. Many groundwater systems have low
to moderate levels of alkalinity but low pH and high DIC values because of the presence
of dissolved carbon dioxide at levels exceeding saturation values. The extent to which the
saturation point is exceeded is critical to determining the increase in pH achievable by
aeration. Achieving pH values greater than 7.0 to 7.5 may not be possible by aeration.
There are a wide variety of aeration systems including diffused bubble systems,
packed or tray tower, and Venturi systems. Any aeration system selected for pH adjust-
ment should be capable of removing at least 80 to 90 percent of the carbon dioxide. Larger
amounts of pH adjustment will require the use of designs that produce higher percentages
of carbon dioxide removal, and pH values above 7.0 to 7.5 may be difficult to achieve by
aeration. If higher values are desired, chemical adjustment may be necessary. One of the
disadvantages associated with aeration is that repumping of the water is required. Some
M58 book.indb 85 11/17/2010 4:25:10 PM

water systems can configure their well, plant, and storage locations to maximize the use
of gravity in the hydraulics of their distribution networks. Some state regulatory agencies
require systems to disinfect the water after aeration so that any microbes introduced dur-
ing aeration will not grow out in the distribution system. Aeration may also oxidize iron
and manganese, resulting in red or black water problems in the distribution system if the
levels of iron and manganese are high enough.
Chemical Inhibitors
The use of corrosion control chemicals is common in the drinking water industry. However,
each chemical has different properties and may not truly inhibit corrosion. Some chemi-
cals simply mask the effects of corrosion to prevent aesthetic problems, such as red or
black water. Currently, the most common inhibitors used in the drinking water industry are
orthophosphates, polyphosphates, blended phosphates, and silicates. The inhibitors can be
classified as sequestering agents, anodic coaters, or cathodic inhibitors (see Figure 4-19).
Except for those chemicals that act as sequestering agents, the chemicals form a protec-
tive, relatively insoluble film on the inside surface of the pipe, acting as a barrier between
the water and the pipe surface. The USEPA also considers chemicals that cause calcium
carbonate scale formation, such as lime, to be corrosion inhibitors. As described in chapter
2, the formation of a calcium carbonate scale should not be considered a form of corrosion
inhibition. Stannous chloride has been used by some utilities but is not currently approved
for corrosion control by the USEPA.
AWWA publishes standards for the various types of corrosion control chemicals. Any
chemical used in water treatment should be approved by NSF International.
When determining inhibitor doses, it is important to consider the seasonal impacts
of water system demand. The Minnesota Department of Health determined that sea-
sonal fluctuations in tap lead and copper concentrations occurred as a result of seasonal
changes in pumping rates (Rezania 2004). It was observed that lead and copper concen-
trations would increase in the summer rather than the winter. It was theorized that the
summer increase was a result of “wearing off” of the protective coating forming on lead
and copper pipe walls. Most utilities add their inhibitor at a constant dose and adjust the
mass feed rate based on flow. Thus, more inhibitor is added in the summer when flows
are high. The pipe wall surface area remains constant regardless of flow and theoretically
would require a constant inhibitor mass feed rate to remain passivated. It was theorized
that the protective coating began to wear off as the flow and total inhibitor mass feed rate
decreased, until the inhibitor mass feed rate reached a low and lead and copper concen-
trations reached a high in the spring and summer. As flows increased and the inhibitor
mass flow rate increased, these protective layers would continue to form. The Minnesota
Department of Health recommended adjusting seasonal inhibitor dosages to account for
the reduction in flow.
SEQUESTERING AGENTS
• Bimetalic Polyphosphates
• Sodium Hexametaphosphate
ANODIC COATERS CATHODIC INHIBITORS

• Sodium Silicates • Calcium Carbonate
• Zinc Orthophosphate • Polyphosphates
• Zinc Orthophosphate
CATHODE
ANODE
METAL
Figure 4-19 Examples of chemical inhibitors
M58 book.indb 86 11/17/2010 4:25:10 PM

The success of an inhibitor is dependant on three factors: dosage, feeding, and flow
rates. The initial dosage of any inhibitor should be two to three times the normal dos-
age of inhibitor to build up a protective film as fast as possible. Building up the film may
take weeks to months depending on the system. The inhibitor should be fed continuously.
Interruptions in feed may cause a breakdown in the scale and prevent reformation of the
protective barrier. The flow rates must be sufficient to continuously transport the inhibi-
tor to all parts of the metal surface; otherwise, an effective protective film will not form.
For example, inhibitors are not effective for controlling corrosion of tanks because the
coating cannot be fed continuously to all parts of the tank surface.
The choice of inhibitor and the necessary dose is dependent on the water system,
operational conditions, distribution water quality, and particular corrosion concerns.
Polyphosphates. Polyphosphates have a successful record of corrosion control in
some waters but have been found to accelerate copper and lead corrosion in most waters.
Dodrill and Edwards (1995) found that for utilities with an alkalinity less than 30 mg/L as
CaCO3 and a pH of 7.4 to 7.8, polyphosphates had an average corrosion by-product release
three to four times lower than utilities using orthophosphates. A study in Wisconsin by
Cantor et al. (2000) looked at the effect of the use of polyphosphate blends and orthophos-
phate on lead and copper pipe corrosion. Their results showed that using a polyphosphate
blend (with polyphosphate contributing two-thirds of the phosphorus concentration and
orthophosphate contributing the other third) increased lead concentrations to four times
what they were in the untreated pipe loop. In general, polyphosphates may be used as a
sequestering agent for iron and manganese but should not be used for lead control. It must
be cautioned that even if polyphosphates are used with the intention of sequestering, lead
and copper may be held by the chemical and increase in the water.
Polyphosphate structures are numerous. Some of the most important structures are
shown in Figure 4-20.
Polyphosphates are best used for the sequestering of metals in a system with low
lead and copper concentrations to prevent red water because polyphosphates stabilize
iron and manganese. Low doses of polyphosphates mask the symptoms of red water as
the metals are tied up as a complex molecule consisting of FePO4 and other metal-inhibitor
compounds. Corrosion rates, however, are not reduced. Thus, it is possible that lead and
copper levels may increase in the water. Polyphosphates should only be considered when
it is known that lead and copper levels would not increase. Polyphosphates also reduce
calcium carbonate scales by the adsorption of the polyphosphate onto the crystal faces,
thus arresting the growth of the crystal.
O− C− O−
__ __ __
O=P O− O=P O− O=P O−
O C O
O=P __ O− O=P __ O− O=P __ O−
O− C− O
n
O=P __ O− O=P __ O−
C− O−
n = 10 to 15
Pyrophosphate ion Tripolyphosphate ion Linear polyphosphate ion
Figure 4-20 Three polyphosphate structures
M58 book.indb 87 11/17/2010 4:25:11 PM

Polyphosphates are cathodic corrosion inhibitors. Divalent cations such as calcium

(Ca+2) at levels of at least 10 ppm are needed along with the polyphosphates for the poly-
phosphates to be effective. Positively charged colloidal complexes are formed that migrate
to the cathode, forming an amorphous polymeric film. This film is self limiting. Analysis
of the film has shown ferric pyrophosphate and iron/calcium metaphosphate to be the
primary constituents. The pH is best maintained at the level of 6.5 to 7.0 because as the pH
drops, the rate of reversion of the polyphosphate to the orthophosphate increases. Typical
dosages are 2 to 4 mg/L. In these applications, the polyphosphate is acting both as a reser-
voir of potential orthophosphate and as a cathodic corrosion inhibitor. More soluble and
less likely to be precipitated in its polymeric form, polyphosphates can be retained longer
in the system than orthophosphates.
Polyphosphate feed points should be separated from the chlorine injection point
by as much distance as possible. The polyphosphate feed point should also be ahead of
the chlorine injection point. If polyphosphate is fed after the chlorine, there is a possibil-
ity that the iron and manganese will be oxidized by the chlorine before the sequestering
action can take place, causing iron and manganese precipitates to be pumped out into the
The feed equipment used for phosphate addition is similar to the equipment used to
feed fluoride, consisting of a storage tank, solution tank, feed pump, and controller to pace
the equipment. The storage tank and the solution tank must contain at least 10 ppm of free
chlorine residual to prevent bacterial growth in the phosphate solution; polyphosphate is
an excellent food source for bacteria.
Solutions of polyphosphates may also be made up from powder in a saturator similar to
the one that is used in making up solutions of dry-fluoride compounds. Phosphate solutions
containing more than 0.5 lb phosphate/gal (60 ppm) may be very viscous. It is important that
any solution be fed within 48 hr of its production. Polyphosphates tend to break down into
orthophosphate, which is much less effective in preventing manganese deposits.
The amount of phosphates required to sequester iron and manganese varies, but in
Minnesota the amount generally has to be approximately two parts actual phosphate (as
product) for one part of iron and manganese to accomplish the desired result. It is impor-
tant to remember that a chlorine residual must be maintained throughout the distribution
system to control bacterial growth. This residual should be greater than 0.2 ppm at the
most distant part of the system.
If the total detention time in the distribution system exceeds 72 hr, the phosphates
may break down and release the iron and manganese in the outer portions of the system. It
is important for operators to know if the detention time will be exceeded and that the iron
or manganese problem may not be resolved with polyphosphate treatment.
In summary, polyphosphates do not have any effect on corrosion inhibition or reduc-
tion of metals release. Polyphosphates are ineffective in stagnant waters. Protection
increases with turbulence. Studies have shown that higher velocities tend to help in cor-
rosion control. Polyphosphates are recommended for sequestering iron and manganese.
However, the distribution system should be carefully monitored to ensure that the product
keeps the metals in solution and that lead and copper have not increased as an unintended
side effect. A higher dose of disinfectant may be needed to prevent microbial growth in
the presence of phosphorus added in the chemical. Polyphosphates should not be used for
control of lead corrosion.
Polyphosphate–orthophosphate blend. Polyphosphate–orthophosphate blends
are available for corrosion control. Polyphosphates are cathodic inhibitors, and orthophos-
phates are anodic corrosion inhibitors. Polyphosphates in solution slowly revert to the
orthophosphate condition. Blends of polyphosphates and orthophosphate can provide
both anodic and cathodic corrosion inhibition; ensure the orthophosphate ion availability
over a longer period of time; and control calcium, iron, and lead deposition. Their primary
M58 book.indb 88 11/17/2010 4:25:11 PM

effectiveness appears to be that the polyphosphate aids the orthophosphate to penetrate

the corrosion scales and form a protective barrier. Orthophosphate–polyphosphate blends
are being produced specifically to use in water systems when an orthophosphate inhibitor
is a viable corrosion-control approach and a polyphosphate is also required to meet other
treatment objectives, such as the control of red water discoloration from iron or hardness
stabilization. Blended orthophosphate–polyphosphates have the potential to provide cor-
rosion control, finished-water stabilization, and distribution system protection. Testing for
both orthophosphates and polyphosphates in the distribution system will assist in deter-
mining the correct inhibitor dosage.
The fraction of orthophosphate in blended formulas ranges from 5 to 70 percent.
Formulas with higher orthophosphate concentrations provide more corrosion control, but
blends with higher polyphosphate concentrations provide enhanced sequestering. Dos-
ages are dependent on pH and calcium concentration of the water. Initial corrosion control
with blended polyphosphates is often slow due to the low dosages used to avoid soft scale
formations and discolored water. Positive corrosion control may take 6 months or longer.
Polyphosphate–orthophosphate blends have been demonstrated to reduce copper
corrosion at rates greater than the rates when orthophosphate is used alone (Edwards
et al. 2001). Application dosages are usually in the range of 0.4 to 1.2 mg/L. The blends
work best in water with a pH range of 7.2 to 7.8.
Orthophosphate. The phosphate-based inhibitors are by far the most commonly
used chemical inhibitors. Two surveys, one in 1994 (Reiber et al. 1997; Dodrill and Edwards
et al. 1995) and one in 2001 (McNeill et al. 2002), examined the use of corrosion inhibitors
by US drinking water utilities. Both surveys found that more than half of all utilities serv-
ing more than 10,000 people use some type of phosphate-based corrosion inhibitor. The
earlier survey found that as the size of utilities decreased, so did the frequency of inhibitor
use.
McNeill et al. (2002) also noted a shift from the use of polyphosphate to orthophos-
phate following implementation of the LCR and noted the potential for polyphosphates to
increase dissolved metal concentrations. The utilities surveyed also cited a variety of rea-
sons for using phosphate-based inhibitors (Figure 4-21). The most commonly cited reason
was lead and copper corrosion control, followed by prevention of red water. But other fac-
tors, such as protection of cement–mortar linings and prevention of calcium scale buildup,
were also noted. The doses reported in the survey ranged from less than 0.2 mg/L as P to
greater than 3.0 mg/L.
Orthophosphates can act as anodic corrosion inhibitors in the presence of oxygen
(e.g., in surface and aerated waters). For example, dissolved oxygen attacking a metal such
as iron forms αFe2O3. The ferrite film gradually spreads, corrosive attack occurring at the
breaks in the oxide film. These gaps are then plugged by the formation of insoluble iron
phosphates resulting from the reaction of sodium orthophosphate with the anodic corro-
sion product. That is,
1. At the anode, Fe0 —> Fe+2 + 2e–

2. Fe+2 + H2PO4 + 2H2O —> FePO4.2H2O + 2H+ + e –
(precipitated)
See chapter 2 regarding the mechanism for corrosion reduction by phosphate inhibi-
tors and chapter 3 for the effectiveness of phosphate inhibitors under certain water quality
conditions.
M58 book.indb 89 11/17/2010 4:25:11 PM

1994 survey 2001 survey

50
45
40
35
Utilities—%
30
25
20
15
10
5
0
te
nd
er
at
at
at
at
ha
th
le
ph
ph
ph
ph
O
sp
B
os
os
os
os
e
ho
at
ph
ph
ph
ph
yp
ph
ho
ho
ly
ho
l
os
Po
Po
rt
rt
rt
ph
O
–o
nc
nc
ho
ly
Zi
Po
Zi
rt
–o
nc
ly
Zi
Po
Type of Inhibitor
Source: Rushing et al. 2002.
Figure 4-21 Type of phosphate inhibitor used by water utilities
1994 survey 2001 survey

90
80 Lead or copper
70
Copper
Lead
60
Utilities—%
50
40
30
20
10
NA NA
0
Prevent Prevent Prevent Prevent Iron Other
Copper or Iron Concrete Calcium or
Lead Corrosion Corrosion Scale Manganese
Corrosion or Red Water Control
Reason for Using Phosphate Inhibitor
Utilities were allowed to select more than one answer, so values do

not total 100%. NA—not applicable; utilities in the 1994 survey
were not asked about the reason for using inhibitors.
Source: Rushing et al. 2002.
Figure 4-22 Reasons for using phosphate inhibitors
Courtesy of Washington Aqueduct.
Figure 4-23 Orthophosphate feed system
M58 book.indb 90 11/17/2010 4:25:26 PM

Orthophosphate is available as straight orthophosphoric acid. There are also neu-

tral orthophosphate products, such as potassium orthophosphate. Orthophosphate is
significantly less expensive than polyphosphates or zinc orthophosphate. Small doses
of orthophosphate appear to significantly reduce iron release, but larger doses around
3 mg/L are required for iron corrosion control.
Orthophosphate works best in a pH range of 7.3 to 7.8. Pipe loop studies have demon-
strated an immediate increase in lead concentrations whenever the pH of a water would
fall out of this range. Normal dosages are 0.5 to 3 mg/L.
Zinc orthophosphate. Zinc orthophosphates were first used for corrosion control
in 1950. Zinc orthophosphates as a technology provide effective and comprehensive corro-
sion inhibition. Traditionally, they have been used to stop colored water caused by bleed-
ing tuberculation and have a long history of success. This technology is very effective in
aggressive waters with no source iron and manganese to sequester and where calcium
stabilization and scaling are not concerns. Zinc orthophosphates have demonstrated
effectiveness in providing lead and copper protection. Zinc orthophosphates also reduce
asbestos fiber counts and protect cement pipe. Most formulations are proprietary and con-
sist of a zinc salt (either chloride or sulfate) combined with orthophosphate, formulated in
a wide variety of zinc-to-phosphate ratios ranging from 2:1 to 1:10. They function by two
distinct mechanisms—film formation and electrochemical passivation. Zinc orthophos-
phate works best in waters with a pH between 7.3 and 7.8. Zinc orthophosphates should
never be fed to a water with a pH above 8.1 because of premature zinc precipitation. In gen-
eral, zinc orthophosphates are less soluble and dissolve less readily than polyphosphates
or orthophosphoric acid. Zinc orthophosphate may perform better at lower dosages than
orthophosphate, particularly in harder waters. The scales of the protected pipes appear to
have equal parts zinc and phosphate, leading to the belief that the zinc and phosphate have
a synergistic effect. It is postulated that the zinc forms carbonate compounds and protects
the cathodic sites while the phosphate forms a protective barrier on the anodic sites.
Recent research, however, has shown that zinc orthophosphate does not demonstrate
any significant advantages compared to straight orthophosphate for the control of lead
and copper corrosion. In a pipe loop study by the DC Water in 2005, zinc orthophosphate
was less effective at controlling lead-oxide corrosion than straight orthophosphate.
Zinc polyphosphates are also available. This technology is appropriate for harder,
less aggressive waters that require some level of sequestering and/or calcium stabilization.
Zinc polyphosphates offer a variety of benefits. In addition to some of the traditional ben-
efits offered by zinc orthophosphates, the polyphosphate component prevents scale and
stabilizes soluble iron, manganese, and calcium by sequestering these ions. Limited lead
protection is obtained through hydrolysis of the polyphosphate to orthophosphate ion.
Products in this category consist of a zinc salt combined with hexametaphosphate in a dry
state. As with polyphosphates, zinc polyphosphates should not be used for lead corrosion.
The fact that zinc orthophosphates and zinc polyphosphates contribute to the zinc
load at the wastewater treatment facility must be considered when selecting a corrosion
inhibitor. The zinc in zinc orthophosphates and zinc polyphosphates may improve corrosion
control in harder waters; however, the zinc-containing products do not appear to offer addi-
tional corrosion control for most waters compared to other orthophosphate compounds.
Silicates. Sodium silicate is available either in a dry chemical form or as a liquid
solution. Silicates are most often used in soft waters with low pH and high dissolved oxy-
gen, although silicates have been proven successful for corrosion control in other waters.
Dosages range from 4 to 30 mg/L. The maximum dosage is 40 mg/L. The concentrations of
calcium, magnesium, chloride, and other materials affect the optimal dosage. Higher dos-
ages are used in waters with higher concentrations of hardness, chloride, and dissolved
solids. It is recommended that an initial startup dose of 24 mg/L be applied. The dose may
M58 book.indb 91 11/17/2010 4:25:26 PM

be incrementally decreased after 60 days to maintain an 8 to 12 mg/L residual. It should be

added at or shortly after oxidant addition.
The protective mechanism of silicates appears to be the formation of a thin layer
over the corroded metal layer as opposed to the mechanism for phosphates, which forms
a protective barrier over the bare metal surface by chemical reactions with the bare metal
surface or corrosion products. The films are self limiting and do not build up in thicker
layers with increased dosages. The films also break down relatively quickly if there is an
interruption in feed.
Silicates are very expensive and cost more than phosphates. In general, silicates
have been found to be less effective for corrosion control than the use of zinc orthophos-
phate or orthophosphate alone. The use of phosphate–silicate blends has increased in the
past few years. The blends appear to be as good as phosphate products alone but are used
at reduced concentrations. It is important to note that more research is required on the
use of silicates to control corrosion. The high costs associated with silicates and the time
required to reduce corrosion in the distribution system have eliminated the use of silicates
as a possible corrosion treatment for many water treatment plants.
Stannous chloride. Stannous chloride has been studied for its potential use for the
control of lead corrosion. The stannous ion is the in-between oxidation state of tin and
may be either oxidized or reduced depending on the atmosphere and the ionic state of the
materials nearby. In theory, stannous ions in the water are drawn to active metal surfaces
that are constantly losing electrons as they oxidize or as they are forced into solution by
electric potentials. The positively charged tin ion is drawn to these negative surfaces and,
depending on conditions, reduces the oxidized ions (rust and corrosion) and also reduces
itself. This action leaves behind an inactive layer of tin and a reducing atmosphere of stan-
nous chloride very close to the surface of the pipe. When an oxidizing agent attacks the tin
surface, the stannous ions protect it, and if the tin surface is diminished, it is replaced by
the ongoing water treatment. Corrosion is still occurring, but the corrosion is now to the
tin layer and not to the pipe material.
Stannous chloride has failed to be a viable corrosion control chemical in numerous
pipe loop studies. Although NSF International approved up to 1 mg/L for use in water, the
USEPA has not approved stannous chloride for use as a corrosion control chemical. Saint
Paul, Minn., is often cited as a case study for the successful use of stannous chloride to
control lead corrosion. Water Research Foundation project 3174 is currently investigating
the use of stannous chloride for control of lead more thoroughly. The project is expected
to be complete in 2010.
Treatment Costs and Dosages

The costs of chemicals used for corrosion control vary. Table 4-8 lists some of the com-
parative costs (2007 data). In addition to the chemical costs, there are several safety
and handling considerations. Table 4-9 lists some relative cost information for different
treatments. For example, lime for pH adjustment is a relatively inexpensive chemical but
requires an investment in feed equipment that may prove maintenance intensive.
Control of Microbially Influenced Corrosion

MIC has been discussed in chapter 2. In order to prevent and control MIC, it is important to
understand the operating conditions of the system. Hence, evaluation of the biological and
abiotic parameters is essential when implementing prevention and control measures.
Methods implemented to prevent MIC should address the following fundamental
issues:
• Inhibition of the growth and/or metabolic activity of microorganisms
M58 book.indb 92 11/17/2010 4:25:26 PM

• Modification of the environment in which the corrosion process takes place to

avoid adaptation of microorganisms to the existing conditions
Techniques for controlling MIC include disinfection, control of water age, flushing,
and nutrient control.
Table 4-8 Comparative cost of treatment chemicals*

Treatment Costs
Chemical Dosage (mg/L) Cost ($/lb) ($/mil gal)
Lime 10–30 0.03 3.00–8.00
Caustic soda 10–30 0.27 13.00–20.00
Soda ash 10–30 0.16 20.00–70.00
Sodium hexametaphosphate 1–4 (PO4) 1.22 10.00–40.00
Bimetallic phosphate 0.5–2 (PO4) 2.05 8.50–35.00
Zinc orthophosphate 0.10–0.5 (Zn) 3.05 3.00–35.00
Sodium silicate 4–10 (SiO2) 0.41 13.00–35.00
Carbon dioxide 5–10 0.07 3.00–5.00
Phosphoric acid 0.5–3 (P) 0.82 3.00–25.00
Monosodium phosphate 0.5–3 (P) 1.60 8.00–50.00
Orthophosphate–polyphosphate blend 0.2–1 (PO4) 3.40 5.00–30.00
*2007 data.
Table 4-9 Relative cost information
Operation and Requirement for

Chemical Installation Costs Chemical Costs Maintenance Process Control
Dolomitic materials High 3–4 times higher Relatively small pH control system to
than calcium avoid high pH leaving
carbonate plant
Calcium hydroxide Limemilk—moderate 2.5 times higher than High pH control; may be
Limewater— calcium carbonate difficult for low
moderately high alkalinity water
Calcium oxide Limemilk—moderate 1.5 times higher than High pH control; may be
Limewater— calcium carbonate difficult for low
moderately high alkalinity water
Sodium bicarbonate Low 40 times higher than Relatively low pH control
calcium carbonate
Sodium carbonate Low 12 times higher than Relatively low pH control
calcium carbonate
Sodium hydroxide Low 20 times higher than Relatively low pH control; may be
calcium carbonate difficult for low
alkalinity water
Sodium silicate Low 5 times higher than Low None
calcium carbonate
M58 book.indb 93 11/17/2010 4:25:26 PM

Secondary Disinfectants
Chlorine. Free chlorine residuals can impact the metal oxidation rates and the
nature of scales that form on the interior of distribution system piping. Chlorine is fre-
quently used in a gaseous form, which hydrolyzes to form hypochlorous acid (HOCl) and
hydrochloric acid:
Cl2 + H2O → HOCl + HCl
Hypochlorous acid is the active species, and its dissociation is pH dependent.
HOCl → H+ + OCl−
At pH 7.5, the concentrations of hypochlorous acid and its ions are equal. At more
alkaline pH values, the equilibrium is displaced in favor of the ionic form. At pH 9.5, all
chlorine is converted to hypochlorite ions, which have poor biocidal activity. Based on
this pH-sensitive equilibrium, the range from pH 6.5 to 7.5 is considered to be ideal for the
action of chlorine for use as a biocide and to control MIC (Videla 2002). Chlorine at lower
pH levels can accelerate corrosion.
Chlorine is an excellent algicidal and bactericidal agent, despite the evidence that the
effective concentration of chlorine is considerably reduced following the penetration of
bacterial biofilms. Measurements performed using selective microelectrodes showed that
the chlorine concentration inside the biofilm represented only 20 percent of the concentra-
tion measured in the bulk liquid in contact with the biofilm. This finding may explain the
decrease in chlorine biocidal efficacy when it is used on biofilms developed by different
species of sulfate-reducing bacteria or by mixed bacterial consortia (Videla 2002).
Other sources of chlorine are salts of hypochlorous acids, such as sodium hypochlo-
rite (NaOCl) or calcium hypochlorite (Ca(OCl)2). Their action is similar to that of gaseous
chlorine. Chlorine dioxide (ClO2) is also used as a substitute for gaseous chlorine because
it does not form hypochloric acid.
The effects of free chlorine on the corrosion of copper in drinking water have been
studied more thoroughly than the effects of other disinfectants because chlorine is a more
common disinfectant used in drinking water distribution systems. Boulay and Edwards
(2001) summarized the relevant results from the 1980s and found that copper corrosion
may be increased or decreased by chlorine depending on pH. At low pH, copper leaching
caused by chlorine was found to be the highest.
Recent research and investigation of the events in Washington, D.C., have observed
that high free chlorine residuals greater than 3 mg/L have been found to inhibit lead cor-
rosion. The use of high free chlorine residuals is not recommended as a corrosion control
technique. Increased chlorine residuals will result in increased DBPs and taste-and-odor
complaints.
Chloramine. Many utilities have changed from free chlorine to chloramine for sec-
ondary disinfection to comply with the Stage 1 D/DBPR. Chloramine is formed when ammo-
nia is bound to chlorine. The chlorine-to-ammonia ratio ranges from 3:1 to 5:1. Because
chloramine is not as strong an oxidant as chlorine, the typical chloramine residuals range
from 2 to 4 mg/L. Chloramine has also been found to be more effective at the higher pHs.
Changes from free chlorine to chloramine for secondary disinfection may, under cer-
tain conditions, also result in an increase in lead and/or copper solubility. The problem
is not that chloramine inherently takes up more metals; rather the problem is the loss
of the protective nature of free chlorine. Chlorine provides a higher OPR, or redox, and
chloramines have a lower ORP. The change in ORP affects the formation of lead scales.
Reiber (2004) reported that a change in disinfectant could impact the redox potential of
the service line surface. The redox potential determines the primary mineral species of
metal that will be present. Changes in redox potential that cause a shift in mineral species
M58 book.indb 94 11/17/2010 4:25:26 PM

from a less soluble to a more soluble species, from Pb4+ to Pb2+ for example, may increase
lead solubility.
Recent Water Research Foundation studies have indicated that most utilities that
switched to using chloramine as a secondary disinfectant did not experience an increase
in corrosion.
Utilities using chloramine must be more vigilant with distribution system main-
tenance and monitoring. Ammonia is a nutrient that may be utilized by AOB resulting in
nitrification. In addition to the potential health effects of high nitrite levels, nitrification
can lead to increased corrosion through microbial activity and lower pH.
Reoptimizaton of Corrosion Control Treatment

Utilities may find it necessary to reoptimize corrosion control treatment whenever
changes are encountered in the source water or when utilities are contemplating treat-
ment changes such as changes to coagulant type or secondary disinfectant. Table 4-9 lists
the impact on corrosion by changes to source of supply, changes to water treatment, and
changes to distribution system activities.
Utilities change treatment processes for a variety of reasons, including
• Changes in source water quality
• Improved cost-effectiveness of existing treatment
• New or more stringent regulatory requirements
• Need to meet or exceed customer expectations (e.g., industrial water user
requirements)
• Aesthetic parameters, such as hardness, color, and taste and odor
The change of a process may have a significant effect on finished water quality param-
eters that impact corrosion. Treatment processes that impact the treated water pH and/or
alkalinity are most likely to cause changes in corrosion chemistry. Lime softening, mem-
brane processes, and ion exchange are three processes likely to impact these parameters.
Other processes that may have negative corrosion impacts include ozonation, granular
activated carbon use, and biofiltration (AWWA 2005).
Changes in distribution system operation practices may impact water quality and the
effectiveness of corrosion control treatment. The operational issues most significant to
maintaining water quality and effective corrosion control are
• Water age
• Blending finished waters of different qualities
• System pressure, direction, and flow of water
Water quality deterioration in a distribution system is often closely associated with
water age within the distribution system. Changes in distribution system operations that
result in significantly higher water age may have negative impacts on corrosion-related
parameters, such as pH (Kirmeyer et al. 2000).
Secondary Effects of Chemical Treatment

Implementation or reoptimization of corrosion control may have secondary impacts.
These potential impacts must be considered when selecting a corrosion control method.
In some cases, a desktop study and baseline monitoring may be inconclusive or there will
be system-specific conditions (e.g., unlined cast-iron pipe, particulates, and changes in
secondary disinfectant) that warrant additional study to determine the impact of changes
on corrosion control treatment effectiveness. The use of blending analysis, treatment
M58 book.indb 95 11/17/2010 4:25:26 PM

Table 4-10 Potential impact of treatment changes on corrosion control
DO = dissolved oxygen
GAC = granular activated carbon
NF/RO = nanofiltration/reverse osmosis
NOM = natural organic matter
TDS = total dissolved solids
Source: American Water Works Association (2005).
M58 book.indb 96 11/17/2010 4:25:29 PM

simulation, loop studies, and increased baseline analysis may be needed to establish opti-
mum water quality and corrosion control conditions. Chapter 5 discusses the types of
monitoring that should be done in the distribution system. Table 4-11 lists the potential of
secondary impacts of corrosion control.
To minimize potential secondary impacts of corrosion control, it is best if a system
makes incremental changes in finished water quality during startup to avoid exposing the
system to large finished water quality changes over a relatively brief time. Rapid changes
in water chemistry may disrupt the system equilibrium and affect the films and scales on
the pipe walls time (Kirmeyer et al. 2000). For systems adjusting pH/alkalinity, a rapid
change would constitute making a 0.1 to 0.2 change in pH units per week in systems with
a high potential for secondary impacts to occur and 0.5 pH units per week in systems with
low potential for secondary impacts (Leitch 1996). For phosphate addition, dosages should
be made incrementally, approximately 0.2 to 0.5 mg/L as PO4. If pH needs to be reduced for
orthophosphate usage, the decrease should occur incrementally over several weeks. To
minimize the risk of adverse impacts, the utility should also avoid making other treatment
changes during corrosion control implementation.
In general, secondary impacts of corrosion treatment may cause increased red water
from scale release of unlined cast-iron or old galvanized steel pipes. Increased total coli-
form and HPC associated with scale release may also occur.
Red water. Red water, also known as yellow or colored water, results from the cor-
rosion of iron-containing pipes used to transport water. Red water may be a secondary
effect when initiating corrosion control. Changes to pH and alkalinity can increase iron
corrosion through the softening of the iron tubercles. The introduction of polyphosphates
or orthophosphates may lead to the release of corrosion by-products. Systems that expe-
rience red water complaints following installation of corrosion control usually have a
concentrated area or large numbers of unlined cast-iron mains. To minimize red water
issues during the initial implementation phase of corrosion control treatment, it is rec-
ommended that an aggressive flushing program be conducted prior to corrosion control
treatment implementation to remove sediment, loose corrosion by-products, and biofilm.
Once corrosion control treatment is implemented, if the alkalinity is greater than 40 mg/L
and stable, DIC greater than 15 mg/L, and the pH is maintained between 7.5 and 9.5, red
water issues should be minimal and may be resolved with flushing.
Bacteria. Increases in total coliform and HPCs may occur following implementation
of corrosion control. A variety of bacterial types in relatively high numbers is found to
be associated with tubercles taken from distribution pipe walls. Corrosion control may
exert a physical or chemical attack on the tubercles, causing them to break from the pipe
wall and releasing the bacteria into the bulk water. Flushing during implementation of
corrosion control will aid in removal of the tubercles and sloughed biofilm, reducing the
duration of increased bacterial levels in the bulk water.
There are concerns that the addition of phosphate-based corrosion control chemi-
cals may cause an increase in bacterial levels because phosphate is a nutrient used by
bacteria. Nevertheless, the use of phosphate-based corrosion inhibitors was found to be
associated with lower coliform levels (LeChevallier et al. 1996). Lowther and Moser (1984)
reported that the levels of coliforms decreased within a few weeks following the applica-
tion of zinc orthophosphate in the Seymour, Ind., system.
M58 book.indb 97 11/17/2010 4:25:29 PM

M58 book.indb 98
Table 4-11 Potential of secondary impacts
Corrosion Postprecipitation Microbiological
Strategy Metals Release of Solids THM HAA Impacts Turbidity Color Taste and Odor
pH increase Decrease in Pb, Precipitation of iron Potential increase Potential Postprecipitation May increase May increase Change in pH may
Cu, Zn. Pos- solids, manganese, decrease of solids may because of post- because of post- affect formation
sible short- and calcium car- create surfaces precipitation of precipitation of of compounds
term increase bonate can occur for microbial solids solids that may resulting in
in alkalinity at elevated pH. growth contribute to taste and odor
could lower color
iron uptake by
increasing buff-
ering capacity.
Iron levels could
be increased if
pH is adjusted
to the point of
minimum buffer
intensity.
PO4 addition Decrease in Pb, Precipitation of alu- No change if No change if Zinc in some for- Reduction in Possible increase Potential growth
Cu, Fe, Zn. minum phosphate pH remains pH remains mulations may tubercles and because of post- of algae in open
Adverse effect can occur if alum constant constant inhibit coliform increased dis- precipitation of reservoirs may
to cement com- is used. Precipi- growth. Possibly infection effi- solids that may affect taste and
pounds if pH tation of calcium improved dis- ciency could contribute to odor
decreases. phosphate can infection effi- reduce turbidity. color
Potential waste- occur at elevated ciency because Postprecipitation
water impacts. pH. Precipitation of reduction of solids may
of zinc carbon- in tubercles contribute to

ate can occur at that provide turbidity.
elevated pH with a growth sur-
the use of zinc face for bacte-
orthophosphate. ria. Potential
increase in bac-
teria if phos-
phate is limiting
factor.
11/17/2010 4:25:29 PM
M58 book.indb 99
Table 4-11 Potential of secondary impacts (continued)
Corrosion Postprecipitation Microbiological
Strategy Metals Release of Solids THM HAA Impacts Turbidity Color Taste and Odor
Polyphosphate Possible increase May reduce precipi- No change if No change if Unknown Turbidity may Color may be Potential growth
addition in Pb, Cu, Fe, tation of iron sol- pH remains pH remains develop at reduced by poly- of algae in open
Zn, Ca ids by acting as a constant constant elevated pH phosphates reservoirs may
chelating agent. with rever- sequester- affect taste and
sion of poly- ing iron and odor
Calcium increase for
phosphate to manganese.
asbestos–cement
orthophosphate.
or cement-lined Reversion of poly-
pipe and softening Turbidity may phosphate to
of cementatious be reduced by orthophosphate
lining. polyphosphate may cause post-
sequester- precipitation of
ing of iron and solids that con-
manganese. tribute to color.
Silicate Possible decrease May prevent iron Potential Potential Potential reduc- Turbidity may be Color may be Elevated pH from
addition in Pb, Cu, Fe, Zn and manganese increase due to decrease due to tion in overall reduced by sili- reduced by sili- use of silicates
precipitation pH increase pH increase corrosion may cate sequester- cates seques- may cause for-
improve dis- ing of iron and tering iron and mation of com-
Zinc carbonate may
infection effi- manganese manganese pounds affecting
precipitate at ele-
ciency because taste and odor
vated pH
of reduced
tubercle for-
mation. Post-
precipitation
of solids may
provide growth
environment for
bacteria.
Source: Kirmeyer et al. 2000.
HAA = haloacetic acid
11/17/2010 4:25:29 PM
Table 4-12 Corrosion control technique selection criteria

Criteria Selection
Effectiveness How effective is each treatment in reducing standing metal concentrations?
How closely will the reductions approach regulatory levels?
How consistent are the standing metal concentrations?
Secondary impacts What secondary impacts are associated with each treatment?
• THM levels
• Disinfection efficiency
• Corrosion by-products
• Iron and manganese precipitation
• Compatibility with existing water quality
• Compatibility with needs of large-volume users or critical industries
• Compatibility with wastewater standards and capabilities
Operability How easy is the treatment technology to use and how safe will it be to operate?
• Liquid versus powder
• Staffing
• Training
• Maintenance
Reliability How reliable is the technology for each treatment approach?
Flexibility How flexible will each treatment be to future water demands? Is the technology
responsive to operational changes?
• Future planning
• Future regulatory issues
Costs How does the present worth value of costs compare between treatments?
• Capital • Inflation
• O & M • Life cycle
• Chemicals • Interest rates
Implementation factors Will there be legal or regulatory ramifications to implementing the treatment?
• Ease of implementation How long will it take to design and construct facilities for each treatment?
• Time to implement How will the public respond to changes in water quality characteristics and any
• Public opinion secondary impacts associated with each treatment?
Wastewater treatment. When selecting a corrosion control treatment, it is impor-

tant to investigate potential issues for local wastewater treatment plants. Many waste-
water treatment plants have limits on the amount of phosphorus they may discharge. If
a drinking water plant is considering the use of a phosphate-based corrosion inhibitor, it
should investigate if the local wastewater treatment plants have the capability to treat the
additional phosphorus loads. Zinc is also a concern for many wastewater treatment plants.
The Clean Water Act limits the discharge of toxic metals such as zinc into receiving waters.
It is possible that the use of zinc orthophosphate may cause excessive zinc loadings, which
the wastewater treatment plant may not be able to treat. The use of zinc-containing cor-
rosion control chemicals may impact the biological processes at wastewater treatment
plants, reducing the effectiveness of wastewater treatment.
Conclusion______________________________________________
Selection of the best corrosion technology is dependent on the intended treatment objec-
tives, a thorough understanding of the characteristics of the water being treated, and the
properties of the technology/products available. In some cases, several different treatment
options will be available. Consideration will also need to be given to impacts of treatment
chemicals on wastewater discharge limits or concentrations of metals in sludge. Tools
such as premise plumbing profiles, pipe loops, and jar testing will assist in determining
M58 book.indb 100 11/17/2010 4:25:29 PM

the appropriate corrosion control technology to employ. Table 4-12 provides a summary of
criteria for selecting a corrosion control technology to evaluate. The complete elimination
of corrosion is almost impossible. Technologies exist to reduce or inhibit corrosion, but
the feasibility of each technology will depend on the economics of an individual system
and the risk/aesthetic tradeoffs. The success of a particular corrosion control treatment
is dependent on the specific water quality and piping of an individual system. A particular
method may be successful in one system and not in another. The most common techniques
for controlling corrosion include distribution system design considerations, water quality
modifications, corrosion inhibitors, and coatings and linings.
References________________________________________________
American Water Works Association (AWWA). Economic and Engineering Services. 1990.
2005. Managing Change and Unintended Lead Control Strategies. Denver, Colo.:
Consequences: Lead and Copper Rule AwwaRF and AWWA.
Corrosion Control Treatment. Denver, Edwards, M., L. McNeill, T. Holm, and M. Law-
Colo.: AWWA. rence. 2001. Role of Phosphate Inhibitors
American Water Works Association Research in Mitigating Lead and Copper Corro-
Foundation and DVGW-Technologiezen- sion. Denver, Colo.: AwwaRF and AWWA.
trum Wasser (AwwaRF and TZW). 1996. Hock, V., E. Kirstein, K. Smothers, and J. Over-
Internal Corrosion of Water Distribution mann. 1999. Laboratory Evaluation of an
Systems. 2nd ed. Denver, Colo.: AwwaRF In-Situ Coating Process for Mitigation
and AWWA. of Lead and Copper in Drinking Water.
Boulay, N., and M. Edwards. 2001. Role of Tem- USACERL Technical Report 99/39. April.
perature, Chlorine, and Organic Matter in Washington, D.C.: US Army Corps of
Copper Corrosion By-Product Release in Engineers.
Soft Water. Wat. Research, 35(3):683–690. Kirmeyer, G.J., G. Pierson, J. Clement, A. Sand-
Cantor, A.F., D. Denig-Chakroff, R.R. Vela, M.G. vig, V. Snoeyink, W. Kriven, and A. Camper.
Oleinik, and D.L. Lynch. 2000. Use of Poly- 2000. Distribution System Water Quality
phosphate in Corrosion Control. Jour. Changes Following Corrosion Control
AWWA, 92(2):95. Strategies. Denver, Colo.: AwwaRF and
Cantor, A., J. Park, and P. Vaiyavatjamai. 2000. AWWA.
Corrosion Control in Small Public Water LeChevallier, M.W., N.J. Welch, and D.B. Smith.
Systems. Midwest Workshop for Small 1996. Full-scale studies of factors related
Public Water Systems. Sponsor: Midwest to coliform regrowth in drinking water.
Technology Assistance Center. St. Louis, Appl. Environ. Microbiol. 62:2201–2211.
Mo. Leitch, R. 1996. Corrosion Control: Begin With
City of Naperville. 2002. Design Manual for pH Adjustment. Water Eng. Management,
Public Improvements. March 2002. 143(1)22–27.
Naperville, Ill.: City of Naperville. Lowther, E.D., and R.H. Moser. 1984. Detecting
Clement, J., M. Hayes, P. Sarin, W.M. Kriven, J. and Eliminating Coliform Regrowth. In
Bebee, K. Jim, M. Becket, V.L. Snoeyink, Proc. of the of the Water Quality Technol-
G.J. Kirmeyer, and G. Pierson. 2002. Devel- ogy Conference. Denver, Colo.: AWWA.
opment of Red Water Control Strategies. Mackintosh G.S., P.F. de Souza, and H.A. de Vil-
Denver, Colo.: AwwaRF and AWWA. liers. 2002. Design and Operation Guide-
DeCarlo, E. 2004. Dielectric Isolation Concepts lines for Municipal Sized Surface Water
in Water Industry Pipeline Corrosion Con- Limestone Contactors. In Proc. of the
trol. ASCE Pipelines. Conference Pro- Biennial Conference of the Water Insti-
ceeding, 146, 78. tute of Southern Africa (WISA). Pub-
Dodrill, D.M., and M. Edwards. 1995. Corrosion lished by Water Research Commission,
Control on the Basis of Utility Experience. Guateng Province South Africa.
Jour. AWWA, 87(7):74–85. Madison, L., G. Gagnon, and J. Eisnor. 2001.
Economic and Engineering Services. 1989. Eco- Corrosion Control Strategies for the Hali-
nomics of Internal Corrosion Control. fax Regional Distribution System. Can.
Denver, Colo.: AwwaRF and AWWA. Jour. Civil Eng., 28(April): 305–313.
M58 book.indb 101 11/17/2010 4:25:29 PM

McNeill, L., and M. Edwards. 2004. Importance Schock, M.R., D.A. Lytle, and J.A. Clement.
of Pb and Cu Particulate Species for Corro- 1995. Effect of pH, DIC, Orthophosphate
sion Control. Jour. Env. Eng., 130(2):136. and Sulfate on Drinking Water Cupro-
National Research Council of the National solvency. EPA/600/R-95/085. Cincinnati,
Academies, Water Science and Tech- Ohio: USEPA, Office of Research and
nology Board (NRC). 2006. Drinking Development.
Water Distribution Systems: Assessing Schock, M.R., I Wagner, and R.J. Oliphant. 1996.
and Reducing Risks. Washington, D.C.: Corrosion and Solubility of Lead in Drink-
National Academies Press. ing Water. In Internal Corrosion of Water
Odegaard H. 1999. Processes for the Removal of Distribution Systems. 2nd ed. Denver,
Humic Substances From Water—An Over- Colo.: AwwaRF and AWWA.
view Based on Norwegian Experiences. Smith, J.E., R.C. Renner, B.A. Hegg, and J.H.
In Conference Proceedings Removal of Bender. 1992. Upgrading Existing or
Humic Substances from Water. NTNU Designing New Drinking Water Treatment
(Norges teknisk-naturvitenskapelige uni- Facilities. Pollution Technology Review,
versitet [Norwegian University of Science Vol. 198
and Technology]). Trondheim, Norway. Spikelmire, W. 2003. Ductile Iron Corrosion
Reiber, S. 2004. Disinfection Byproducts vs. Factors to Consider and Why. ASCE Pipe-
Corrosion: A Case Study on the DC Water lines. Washington D.C.: ASCE.
Lead Experience. Presentation to USEPA, Sung, W. 2004. Corrosion Control and Chloram-
April 2004. ination: Discolored Water and Nitrifica-
Reiber, S., S. Edwards, S. Poulson, S. Perry, tion. In Proc. of the World Water Congress.
S. Patel, and D. Dodrill. 1997. A General Washington D.C.: American Society of
Framework for Corrosion Control Based Civil Engineers.
on Utility Experience. Denver, Colo.: Trussell, R.R., and I. Wagner. 1996. Corrosion
AwwaRF and AWWA. of Galvanized Pipe. In Internal Corrosion
Rezania, L. 2004. Optimizing Phosphate Treat- of Water Distribution Systems. 2nd ed.
ment to Minimize Lead/Copper Seasonal Denver, Colo.: AwwaRF and AWWA.
Variations. Waterline, Minnesota Depart- USEPA. 1984. Corrosion Manual for Internal
ment of Health, 12(3):2. Corrosion of Water Distribution Sys-
Rompre, A., M. Prevost, J. Coallier, P. Brisebois, tems. EPA/570/9-84-001. Washington, D.C:
and J. Lavoie. 2000. Impacts of Implement- USEPA, Office of Water.
ing a Corrosion Control Strategy on Bio- USEPA. 1986. Limestone Bed Contactors for
film. Water Sci. Tech., 41(5). Control of Corrosion at Small Water
Rushing, J., L. McNeill, and M. Edwards. Utilities. EPA/600/2-86/099. Cincinnati,
2002. Some Effects of Aqueous Silica on Ohio: USEPA.
the Corrosion of Iron. Wat. Research, USEPA. 2003. Revised Guidance Manual for
37(5):1080–1090. Selecting Lead and Copper Control Strat-
Schock, M. 1989. Understanding Corrosion egies. EPA/816-R-03-001. Washington,
Control Strategies for Lead. Jour. AWWA, D.C.: USEPA, Office of Water.
81(7):88. Videla, H. 2002. Prevention and Control of Bio-
Schock, M. 1999. Internal Corrosion and Depo- corrosion. International Biodeteriora-
sition Control. In Water Quality and tion and Degradation, 49:259–270.
Treatment. 5th ed. Denver, Colo.: AWWA. Volk, C., E. Dundore, J. Schiermann, and M.
Schock, M., and R. Giani. 2004. Oxidant/Disin- LeChavallier. 2000. Practical Evaluation
fectant Chemistry and Impacts on Lead of Iron Corrosion Control in a Drinking
Corrosion. In Proc. of the AWWA Water Water Distribution System. Wat. Research,
Quality Technology Conference. Denver, 34(6):1967–1974.
Colo.: AWWA.
M58 book.indb 102 11/17/2010 4:25:29 PM

AWWA MANUAL M58
Chapter 5
Implementing Corrosion
Control Treatment
Richard Giani
DC Water
Christopher P. Hill
Malcolm Pirnie Inc.
Elizabeth Turner
City of Dallas, Texas
Steve Reiber
HDR Inc.
Abigail F. Cantor, P.E.

Process Research Solutions LLC
Introduction_____________________________________________
Achieving corrosion optimization is extremely important to every drinking water utility,
including drinking water systems that may not be currently applying corrosion treatment.
Normally corrosion control is associated with meeting the USEPA LCR or minimizing lead
and copper at the customer tap. However, corrosion control optimization also encompasses
many common phenomena such as iron solubility, precipitation from unlined cast-iron
mains, and cement dissolution from newly lined pipes, which might leach other metals or
affect localized water quality. It could also include precipitate formation in the distribution
system such as calcium carbonate, aluminum phosphates, or ferric and manganese precip-
itates, which can cause buildup of particulates, decrease corrosion treatment efficiency,
decrease water pressure, and increase cloudy or discolored water at the customer tap.
103
M58 book.indb 103 11/17/2010 4:25:30 PM

• When initiating corrosion control treatment or changing treatment or sources that can affect metal solubility,
bench testing, pipe loops, and pilot studies are preliminary tools that can be utilized to gather information
relating to treatment effectiveness and unintended consequences.
• Bench tests can predict chemical dosages, changes in water quality parameters, and calcium and phosphate
precipitation.
• Pipe loops can be used to determine metals release from existing pipe scales and new pipe.
• Coupon rigs and electrochemistry (EC) loops can provide supplemental pipe loop information relating to
corrosion rates.
• Coupon techniques are the original quantitative measure for assessing corrosion rates. Coupon tests give
a positive indication of the extent of metal oxidation (the mass of metal lost) as well as visible evidence of
corrosion morphology, such as pitting and scale mineralogy.
• When using coupons, the highest corrosion rates occur at the beginning of the exposure process and then
rapidly decrease to a lesser and more constant rate.
• Process Research Solutions (PRS) monitoring stations are similar to pipe loops but have several more
advantages for use in routine distribution system monitoring and off-line chemical comparison tests. Pipe film
analyses on the metal plates from the stations can be performed easily, giving information about older existing
pipe films in the distribution system.
• EC monitoring can provide a snapshot in time of the corrosion process on a specific pipe specimen under a
given water quality condition. It is, effectively, the only technique that allows this snapshot.
• EC techniques are attractive because of their speed and flexibility; however, EC measures the corrosion process
on the base metal of the plumbing material, not the rate at which metal is actually released from the corrosion
scale to the water in contact with the plumbing surface.
• Premise plumbing profiles are useful to determine metal leaching from the residential pipe scales, the stability
of the pipe scales (i.e., how readily is the scale breaking off of the pipe wall), the concentration and types
of particulates disassociating from the scale, and microbiological activity with respect to nitrification and
localized corrosion.
This chapter will introduce corrosion assessment tools and apply them to the factors
discussed in chapters 3 and 4 with the goal to initiate a proactive corrosion control treat-
ment program.
Corrosion control programs should achieve the following goals:
• Minimize metals release
• Maintain biological and chemical stability
• Avoid unintended consequences
To ensure these goals are met, utilities that are adding corrosion treatment for the
first time; diagnosing corrosion treatment due to increased corrosion activity; or chang-
ing any treatment, dosages, or source water that can have direct or secondary impacts on
water quality parameters should use some or all of the assessment tools described in this
chapter prior to implementing full-scale treatment operations to ensure the listed goals
are met.
The intended goal of this chapter is to provide several options in each step that a
utility can select from that will best fit its need in achieving a successful corrosion control
program (Table 5-1).
M58 book.indb 104 11/17/2010 4:25:30 PM

Implementing Corrosion Control Treatment 105
Corrosion Indexes_______________________________________
As discussed in chapter 2, most corrosion indexes are based on calcium carbonate or
carbonate hardness’ saturation point. In general, a positive number indicates the potential
for calcium or magnesium constituents to precipitate. In the past, this was thought to be
helpful in that a positive index was associated with precipitation and therefore no disso-
lution of metals from the pipe wall (i.e., no corrosion). However, as discussed in chapter
2, calcium precipitation does not form uniform corrosion protection but may cause pipe
stress due to overprecipitation in certain areas of the distribution system. Indexes such
as the Langelier Index (LI) and the Calcium Carbonate Precipitation Potential (CCPP) can
help facilities avoid calcium carbonate precipitation when identifying optimal pH ranges
for corrosion control.
Bench Testing_ ___________________________________________

When preparing to initiate a corrosion control program or a new treatment change, bench
testing is an easy process that can provide general yet pertinent information related to
water quality parameter changes. Bench tests can
• Provide information related to anticipated chemical dosages such as the dosage of
a pH adjustment chemical needed to obtain a target pH
• Reveal potential unintended consequences (excessive CaCO3 precipitation, signifi-
cant ORP change, phosphate reduction, and so on)
• Assist in deciding if additional water quality parameters should be monitored
• Provide design parameters for a pilot study
Many types of bench tests can be conducted depending on plant treatment schemes,
chemical addition, baseline water quality, and so on. However, for corrosion purposes, a
utility could conduct a specific test listed for the following scenarios:
• Calcium carbonate precipitation test: Adjustment of pH, alkalinity, or calcium
hardness
• ORP test: Change in disinfectant or disinfectant concentration
• Coagulation test for chlorides:sulfate: Change in coagulant or coagulant
concentration
• Phosphate test: Addition of a phosphate-based inhibitor
Calcium Carbonate Precipitation Test Procedure

A calcium carbonate precipitation test will provide an understanding of how much cal-
cium carbonate could precipitate in the distribution at the target pH. This test is designed
to help utilities avoid excessive precipitation and should be conducted if computer models
during the desktop study predict a positive or near-positive calcium carbonate precipita-
tion potential (CCPP) at the newly targeted pH or alkalinity. This type of test will measure
the true CCPP under real water quality conditions.
To conduct a CaCO3 precipitation simulation, follow these steps.
1. Collect water from a source that is to undergo pH or alkalinity changes. This
water is referred to as baseline water.
2. In a 1-L beaker, raise the pH and/or alkalinity of the baseline water to the required
level using the desired pH adjustment chemical. Once achieved, continue slowly
stirring the water in the beaker.
M58 book.indb 105 11/17/2010 4:25:30 PM

3. After 24 hr of mixing, remove a sufficient volume of water from the jar to analyze
for either total calcium or total calcium hardness.
4. Remove another 50 to 100 mL of jar test water and filter through a 0.45- μm filter.
5. Measure the calcium or calcium hardness of the filtered water, which are the dis-
solved calcium in the water.
6. To obtain the amount of calcium that has precipitated, subtract the filtered or dis-
solved calcium from the unfiltered total calcium level.
If elevated levels of magnesium exist in the water, replacing calcium hardness with
total hardness measurements will account for both calcium and magnesium precipitates.
Theoretically, precipitate concentrations greater than 10 mg/L as CaCO3 (>4 mg/L as Ca)
are undesirable because they may lead to hydraulic stress in distribution mains, reduce
the life of residential hot-water heaters, and interfere with passivating pipe scales. Maxi-
mum precipitate levels may vary between utilities. The city of Tampa Bay established a
lower maximum calcium precipitation concentration of 7 mg/L as CaCO3 in the distribu-
tion system (Brodeur and Flourence 2007). At higher levels, the city was observing a high
number of service pump failures and inaccuracies in residential meters.
It is important to note that groundwater systems may contain elevated carbon diox-
ide levels compared to surface water systems. When running bench tests, open beakers
are acceptable for surface water sources and groundwater sources that are exposed to the
atmosphere prior to pH adjustment. However, groundwater systems that do not contain
atmospheric exposure (i.e., hydropneumatic tanks—see Figure 5-1) prior to pH adjust-
ment should use closed containers such as biochemical oxygen demand bottles in order to
prevent the gas from escaping. Carbon dioxide (carbonic acid) may be a factor in predict-
ing calcium carbonate precipitation in groundwater systems where pH adjustment chemi-
cal is injected into the line before the water is exposed to atmospheric pressures.
Carbon dioxide (carbonic acid in water) will convert into carbonate as the pH
increases. This conversion requires more pH chemical and increases alkalinity and DIC
as the pH continues to increase. In this situation, calcium carbonate may precipitate at a
lower pH than it would under atmospheric conditions.
CCPP can change in surface water systems on a seasonal basis. Therefore, bench test-
ing should be conducted under source water conditions containing the highest DIC levels.
pH Adjustment
Well Chemical
To Storage or
Distribution System
Courtesy of DC Water.
Figure 5-1 Diagram of pH adjustment chemical injected into line before water from hydropneumatic
tank is exposed to atmospheric pressure
M58 book.indb 106 11/17/2010 4:25:31 PM

Figure 5-2 Apparatus setup for filtering calcium carbonate and other precipitates such as iron,
manganese, and phosphate
As an example of a calcium carbonate precipitation bench test, consider a surface

water utility that needs to raise the entry-point pH from 7.5 to 8.3. In addition, the cur-
rent soda ash feed will be replaced with caustic soda. Desktop models predict a CCPP
of 1 mg/L as CaCO3. The operator conducts a bench test to verify the model by vigor-
ously mixing caustic soda to filter effluent water until the pH increases to 8.3. Once the
target pH is reached, mixing is reduced to a very slow velocity for 24 hr. After mixing is
complete, the operator measures the amount of calcium hardness in the jar and obtains a
reading of 100 mg/L as CaCO3. Approximately 100 mL of water is filtered through a 0.45-µ
filter (Figure 5-2). The operator proceeds to measure the calcium hardness level of the fil-
tered water and obtains a reading of 96 mg/L as CaCO3. Subtracting the dissolved calcium
from the total calcium (100 mg/L as CaCO3 – 96 mg/L as CaCO3), a reading of 4 mg/L as
CaCO3 is obtained. Results were higher than predicted by the model; however, the precipi-
tation potential should not cause a reduction in hydraulics, although calcium carbonate is
expected to moderately accumulate in dead-end or low-flow areas. The utility may want to
further study dead-end precipitation effects by piloting an isolated section of the distribu-
tion system.
Oxidation–Reduction Potential Test

The ORP test should be conducted when a utility is changing disinfectants or changing
concentrations of the same disinfectant. The purpose of the test is to measure current oxi-
dation potentials and compare them to the oxidation potentials that occur due to a change
in concentration of the current disinfectant or a change to an entirely different disinfec-
tant. Changes in ORPs greater than 100 mV may warrant additional studies or construction
of pipe loops in order to evaluate the full effect of the disinfectant change.
To run an ORP test, fill one, 1-L beaker with finished water (entry point) and another
1-L beaker with baseline water (baseline refers to treated water collected immediately
before secondary disinfection is added or collected immediately before primary disinfec-
tion is added if there is no secondary disinfection treatment). To the baseline beaker, add
the new disinfectant at the designated concentration; or if the same disinfectant, add the
new concentration. Using an ORP meter, measure ORP in each beaker. It should be noted
that many pH meters have an option to measure ORP using a standard ORP probe.
M58 book.indb 107 11/17/2010 4:25:32 PM

Coagulation Test for Chlorides:Sulfate

A change in coagulant, specifically a change from a sulfate-based coagulant to a chloride-
based coagulant, will cause a change in the chlorides-to-sulfate ratio. When changing
coagulant chemicals or if major changes in current coagulant chemical concentrations
(usually due to enhanced coagulation) occur, the utility should conduct a bench test to
determine chloride and sulfate concentrations at the filter effluent.
To conduct this test, run a standard jar test using several 2-L square jars with the cur-
rent coagulant and obtain the optimal concentration that will allow for best settling. Using
clean ½-in. tubing, siphon the top 200 mL of the jar that displayed the best settling. Filter
the siphoned water through a 0.45-µm filter and measure the chloride and sulfate concen-
trations from the filtered water. These concentrations represent the amount of dissolved
sulfate and chloride that can pass through the filtration process. Repeat the procedure for
the new coagulant. Determine the chlorides-to-sulfate ratio. Ratios above 0.6 mg/L may
warrant pipe loop analysis or pilot testing.
Phosphate Test
A surface water system planning to add a phosphate-based chemical to the water may
want to conduct bench tests in order to determine the amount of reactive phosphate avail-
able for passivation. Phosphate will react with dissolved aluminum and iron to form phos-
phate precipitates. This reaction is common when improper coagulation occurs and the
dissolved metals pass through the filter. In addition, water with saturated phosphate levels
(>3.0 mg/L as PO4) can form precipitates with ferrous and calcium compounds (Tesfai
et al. 2006) found in the distribution system. If using a polyphosphate or blended phos-
phate, bench tests might provide information relating to the breakdown of polyphosphate
to orthophosphate. If phosphates are used for sequestration and corrosion purposes,
bench tests can also identify time-release of iron and manganese.
Steps for running a phosphate bench test are as follows:
1. Using finished water, add the desired phosphate concentrations to a 1-L jar.
2. Mix rapidly for 30 sec and then continue to mix at a slow rate for 24 to 48 hr or to
the longest detention time in the system.
3. Measure orthophosphate.
4. Filter approximately 100 mL of sample through a 0.45- µm filter
5. Measure the filtered orthophosphate level. This amount is the remaining phos-
phate that is available for corrosion control.
6. Subtract the filtered orthophosphate concentration from the unfiltered orthophos-
phate concentration. This amount is the precipitated phosphate. Precipitated
phosphate levels greater than 1.0 mg/L as PO4 might warrant pilot testing in the
7. If using blended or polyphosphate for sequestration purposes, orthophosphate
readings should be taken at the start of the test as well as at the end. An increase
in unfiltered orthophosphate will reflect the amount of polyphosphate that was
converted to orthophosphate. In addition, measurements for iron and/or manga-
nese should be included in steps 3 through 6. When polyphosphate breaks down, it
can release bound iron or manganese, which when exposed to a disinfectant will
become an oxidize precipitate. Precipitated iron levels greater than 0.3 mg/L and
precipitated manganese levels greater than 0.05 mg/L may warrant pilot testing.
Bench tests can provide purposeful information relating to potential reactions with
baseline water when adding a new chemical. However, they cannot provide information
relating to how the water reacts with pipe scales such as metals release, scale stability,
M58 book.indb 108 11/17/2010 4:25:32 PM

and biofilm formation. Examination of pipe scales and conducting pipe loop operations
allows utilities to monitor these conditions in a controlled atmosphere.
Examination of Pipe scales_______________________________

Conducting XRD and SEM analysis on pipe wall scales can provide supplemental infor-
mation regarding existing scale formation. This data can provide a picture of historical
effects from past treatment and water quality effects on pipe walls. The data can also pro-
vide insight into how a scale might react to a change in drinking water treatment or source
water change. For example, in the District of Columbia, when lead compliance samples
were exceeding USEPA action levels, XRD samples from lead service lines removed from
residents’ homes revealed thick scales of lead oxide (PbO2). This finding helped experts
understand the structure of the scale and its lead-leaching potential into the drinking
water. Since the introduction of orthophosphate as treatment to the lead-oxide scale, XRD
samples have displayed slow changes of the lead oxide to lead phosphate, providing infor-
mation to the utility that the scale is becoming more stable and that the lead is transition-
ing into a form that is less soluble.
XRD is a procedure where X-rays are scattered by crystal atoms on the pipe wall, pro-
ducing a diffraction pattern that yields information about the structure of the crystal. To
conduct an XRD for corrosion pipe-surface information, typically 1-ft sections of service
lines or internal copper plumbing pipes are removed from the system. Pipe specimens are
cut longitudinally with a band saw having a fine metal-cutting blade (Schock and Giani
2004) (Figure 5-3). Scales are then carefully removed to conduct analysis. Figure 5-4 illus-
trates XRD patterns for lead-oxide scales.
When removing the pipes from the distribution system, bending the pipes should be
avoided. Bending pipes will dislodge the scale. Pipe scales should remain moist once they
are removed from the distribution system until they are cut at the lab. Dry scales are frag-
ile and may dislodge during shipment to the laboratory. Sealing the pipe ends with plastic
wrap will maintain the moisture (Figure 5-5). To prevent any damage, it is important that
the pipe be well protected during delivery to the laboratory and that dropping and mis-
handling be avoided.
Courtesy of USEPA.
Figure 5-3 Pipes with lead-dioxide scales cut horizontally prior to XRD
M58 book.indb 109 11/17/2010 4:25:35 PM

DCUNK1 L1 00-013-0131> Hydrocerussite, syn - Pb3(CO3)2(OH)2

DCHENW2 L1 00-046-1045> Quartz, syn - SiO2
DCHENW1 L1 00-025-0447> Plattnerite - PbO2
DCMONW! L1 00-011-0549> Scrutinyite, syn - PbO2
1500
1000
500
Intensity, Counts
0
10 20 30 40
1500
1000
500
0
50 60 70 80 90
Two-Theta, deg
Courtesy of USEPA.
Figure 5-4 Lead-speciated XRD patterns
Courtesy of USEPA.
Figure 5-5 Example of a capped pipe ready for transportation
Pipe loops________________________________________________
Another alternative to collecting metal release data is to build pipe loops using existing
lead pipes from customer homes or service lines. Pipe loops may also include water ser-
vice lines or small mains removed from the ground. Lead and iron pipes removed from
the distribution system are recommended as opposed to new pipes in order to accurately
portray the exact scale buildup on the internal pipe walls. Copper pipe loops could contain
both new and existing pipe material. Newer copper pipes will tend to leach more copper
until passivated, and older pipes may contain existing biofilms and mineral scales that
have deposited over time.
During operations, loops could run on a 24-hr basis, thus simulating water traveling
through a main. The pipe loop may also stagnate for several hours, simulating residential
scenarios or lead and copper compliance requirements (stagnate for 6 hr).
M58 book.indb 110 11/17/2010 4:25:39 PM

Pipe-loop leaching data can be collected on a daily basis in an environment that is

more accessible than collecting samples in customer homes. Also, pipe loops placed in the
distribution system may be more representative of capturing all the reactions that occur
in the distribution before reaching customer premise plumbing.
There are two general types of pipe loops: circulation loops and flow-through loops.
Circulation Loops
Circulation loops are closed-loop systems in which water is brought into the loop and
then allowed to circulate or remain in the system for a certain timeframe (normally 24 hr).
Figure 5-6 contains a general schematic of a circulation-loop system. Figure 5-7 shows an
example of an actual closed-loop system. When designing the pipe loop, make sure the
pipe volume of each pipe material is equivalent to conduct a full assay of required water
quality parameters. From 9 ft of 5/8-in. inside-diameter lead pipe, 500 mL of water can be
obtained.
In closed-loop systems, water is poured into the holding reservoir. Typical pipe-loop
reservoirs vary from 5 to 50 gal in volume. In order to obtain leaching rates of current
conditions, it is recommended to use fresh distribution water changed out on a daily basis.
Baseline data should be collected from the pipe-loop reservoir before the circulation pro-
cess begins each day. Baseline data should consist of typical water quality parameters
collected in the distribution system (pH, alkalinity, phosphate, and so on) and those that
may change during the circulation period (chlorine, pH, lead, copper, and so on).
Once the circulation process is initiated and the water has time to mix in the tanks,
the loops should be shut down after 30 min of circulation to allow the water to stagnate
for 6 to 8 hr if monitoring for lead and copper solubility levels. At the end of the stagna-
tion period, samples are collected at a tap at the end of the pipe gallery and analyzed for
specified parameters discussed in chapter 3. Another sample event should occur at the end
of the circulation cycle and could look at changes in water quality parameters over 24 hr,
such as chlorine loss, pH changes, and lead and copper accumulations.
Reference Multi - channel

Multi-Channel
Multi Channel Meter
Electrode Voltmeter
3- ft PbPb
foot pipe
pipe
¾-in.
3/4”I.D.
I.D.
Sampling Data
Outlet Acquisition
100-L
Recirculation
Tank
Drain
Courtesy of Hatch Mott McDonald.
Figure 5-6 Circulating-loop schematic
M58 book.indb 111 11/17/2010 4:25:40 PM

Figure 5-7 Circulating loop with lead service lines
100
90
80
70
Pipe Loop
Stabilized
60
Lead, ppb
50
40
30
20
10
0
Thu 11/09 Fri 12/29 Sat 02/17 Sun 04/08 Mon 05/28
Date
Figure 5-8 Lead release during pipe-loop conditioning from pipe containing lead-oxide scales
M58 book.indb 112 11/17/2010 4:25:41 PM

When conducting a loop study, particularly when using harvested pipe segments, it is
important to exercise extreme care so as not to disturb existing pipe scales and important
to run the loop at baseline conditions for sufficient time to allow the pipe scale to stabi-
lize (which can take up to several months). After stabilization, metals release will vary
slightly. Figure 5-8 graphs the conditioning period of a pipe loop using lead release data.
Lead or metals release will stabilize, indicating conditioning is complete. The particular
circulation loop being graphed in Figure 5-8 took almost 3 months to stabilize. Original
pipe scales consisted primarily of lead phosphate. Corrosion treatment in this loop was
orthophosphate (3.5 mg/L).
Water exposed to the pipes should be representative of the new water quality once
the loops are conditioned. When adding new treated water, it is important to maintain
chemical stability as much as possible in order to obtain representative metals release.
In closed-loop systems, fresh treated water should be poured into the reservoirs on a
daily basis. Thomas et al. (2004) monitored lead saturation in a circulated pipe loop during
stagnation (Figure 5-9) and determined that lead release in a closed system will decrease
over time, provided water quality conditions remain the same. Therefore, water circulat-
ing in excess of 24 hr may not provide accurate metal solubility. Baseline data should be
collected from the pipe loop reservoir before the circulation process begins and should
consist of typical water quality parameters collected in the distribution system (pH, alka-
linity, phosphate, and so on) and those that can change during the circulation period (e.g.,
chlorine, and so on).
Disinfectant may have to be added at higher concentrations or more frequently dur-
ing the 24-hr cycle if residual is lost. Disinfectants such as chlorine can react with the pipe
scale, oxidizing the metals and neutralizing microbes. Once residuals are lost, the reaction
ceases. For example, chlorine added to a lead circulation pipe loop at a concentration of
3.0 mg/L may generate an ORP of 650 mV. Within several hours, all the chlorine reacts with
the lead, thus reducing the ORP level of the water to 450 mV. In normal household situa-
tions, internal pipes are exposed to fresh water containing chlorine residual several times
during a 24-hr period, thus the lead–chlorine reaction cycle occurs each time the water is
refreshed, but in the circulation loop it only occurs once.
200
150
Lead concentration, ppb
100
50
2
R = 0.955
Kuch-Wagner Model
Observed
0
0.1 1.0 10.0 100.0
Stagnation time, hr
Courtesy of Hatch Mott MacDonald.
Figure 5-9 Metals release over time in a stagnated lead pipe loop
M58 book.indb 113 11/17/2010 4:25:41 PM

Flow-Through Pipe Loops

Flow-through pipe loops are usually more accurate than circulation loops because the
water flowing through the loops is always “fresh” and is more representative of changes
in water variability during a 24-hr period. One of the few disadvantages of flow-through
loops is that they use much more water than circulation loops use and require more main-
tenance, especially when having to add chemical treatment using chemical feed pumps
(Figures 5-10 and 5-11).
When attempting to monitor microbiological activity in the distribution system, flow-
through loops are much more accurate compared to circulation loops. In flow-through
loops, disinfectants are always “fresh” when entering the loop system.
Sampling events in flow-through loops are similar to events described in the circulation-
loop system. When monitoring for lead and copper leaching, a daily stagnation period of 6
to 8 hr is recommended.
Figure 5-12 displays a pipe-loop graph illustrating how changes in disinfectant (alter-
nating chlorine and chloramine) have an affect on lead (IV) release. Each time the disin-
fectant is switched, lead concentrations change. For example, lead (IV) concentrations in
the loop decrease when switching from chloramine to chlorine and slightly increase when
reverting back to chloramine. A properly conditioned pipe loop will allow utilities to iden-
tify the slightest changes in metals release. That is, once a pipe loop becomes conditioned,
slight changes in metals release become apparent. It was also noticed that for the pipe loop
graphed in Figure 5-12, increasing water temperature had a slight compounding impact.
Courtesy of the US Army Corps of Engineer’s Washington Aqueduct.
Figure 5-10 Flow-through system, view 1
M58 book.indb 114 11/17/2010 4:25:42 PM

Figure 5-11 Flow-through system, view 2
WA Pipe Loops
Dissolved Lead and Temperature vs. Time - Rack 3 - Loop 3C
35 25
Switched from Chloramine to Chlorine Switched from
Chlorine
to Chloramine
30
Switched from
20
Chloramine
to Free Chlorine
25
Dissolved Lead, ug/L

15
Temperature
20
15
10
10
5
5
0 0
2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/05 11/24/05 1/3/06 2/12/06 3/24/06 5/3/06
Temperature Dissolved Lead Date
Figure 5-12 Data from a pipe loop that displays changes in lead (IV) release as the loop was
alternated with chlorine and chloramine over time
Coupon Studies___________________________________________
Coupon studies are the original quantitative measure for assessing corrosion rates and for
certain metal types, they are still a preferred measure. The basis for all coupon protocols
is the sustained exposure of a well-defined metal specimen (coupon) of known geometry
M58 book.indb 115 11/17/2010 4:25:44 PM

and mass to a given flow and water chemistry. Of the different analytical methodologies,
long-term coupon weight loss is one of the most useful and readily interpretable corrosion
rate measures. Coupon tests give a positive indication of the extent of metal oxidation (the
mass of metal lost) as well as visible evidence of corrosion morphology (pitting depth and
area). They are useful in the analysis of scale adhesion and mineralogy and are applicable
to many types of metal plumbing surfaces. Coupon exposures are often the referee mea-
surement against which other forms of analysis are calibrated and compared.
Important coupon testing criteria include the following:
1. The coupon metal must be representative of the piping material of interest.
2. The water chemistry of the coupon exposure must be indicative of the distribu-
tion system.
3. The hydraulic regime across the coupon must reflect the flow regime across the
distribution system pipe wall or plumbing appurtenance.
4. The duration of the test must allow for development of corrosion scales and/or
passivating films that influence the corrosion rate of the underlying metal.
Coupon data are commonly presented as a penetration rate expressed in terms of
unit area of mass loss per day of exposure (g-metal/cm2/day). It is common, although not
necessarily correct, to assume the metal loss is evenly distributed over the entire coupon
test surface, making it possible to present the data as a uniform surface penetration rate.
On metals that generally corrode uniformly (i.e., copper, copper alloys, Pb/Sn solders, and
so on), it is acceptable to assume uniformity of metal loss and that uniformity be used to
calculate an approximation of service life. This type of analysis may, however, be inap-
propriate for pitting surfaces or other localized forms of corrosion, where a uniform rep-
resentation of metal loss is likely to grossly overestimate remaining service life. In such
cases, penetration of the pipe wall will occur in a small fraction of the time predicted by a
uniform assumption of corrosion.
As a direct measure of corrosion rate in a drinking water environment, coupon tech-
niques are sometimes limited to those metals, such as iron and steel alloys, where the
nominal corrosion rate produces sufficient mass loss to be readily measured. Under some
exposure conditions, metals such as copper, copper alloys, and lead have such a low nomi-
nal corrosion rate that the mass loss cannot be precisely measured unless extended expo-
sure periods are employed.
There is no single standard regarding coupon geometry, materials, or exposure pro-
tocols in drinking water systems. While some coupon techniques have been developed
specifically for drinking water distribution systems, others have been borrowed from dif-
ferent industries. The ASTM has certified several methodological variants for use in the
evaluation of metal loss on coupon exposures. The most widely used technique (ASTM
D2688-83 Method B, 1983a) relies on flat, rectangular coupon specimens mounted on non-
metallic stems and inserted directly into the flowstream of the pipe, usually at an elbow or
tee (see Figures 5-13 and 5-14). This technique can be used to make relative assessments
of corrosion at different locations in a distribution system and can be used for compara-
tive analyses of corrosion inhibitors. However, because the hydraulic flow lines around a
flat coupon positioned midpipe are substantially different from the flow lines at the pipe
wall, this coupon technique may be inappropriate when a precise estimate of piping cor-
rosion rates is required. Also, it is frequently difficult to obtain flat coupons that are truly
representative of pipe materials.
Table 5-2 presents a comparative summary of published coupon protocols, including
geometry and exposure conditions.
M58 book.indb 116 11/17/2010 4:25:44 PM

Photo courtesy of Metal Samples Company.
Figure 5-13 Typical metal coupons
Figure 5-14 Typical coupon study apparatus
A rigorous coupon evaluation involves measuring weight loss over an extended

period of time and requires coupon sacrifices at multiple points in the exposure cycle. A
weight loss against time curve is drawn, and the corrosion rate at any point is the gradient
of that curve at that exposure duration. Experience has shown that on most metal sur-
faces, corrosion rates change over the course of the exposure, with the highest corrosion
rates occurring at the beginning of the exposure and then rapidly decreasing to a lesser
and more constant rate. Hence, any comparison of corrosion rates for a particular metal
must be standardized to a specific exposure duration.
PRS Monitoring Stations_ ________________________________

An alternative to pipe loops are PRS monitoring stations. These stations can be used for
off-line corrosion control chemical comparison tests and for continuous monitoring of
water quality at critical locations in the water distribution system. The standard configu-
ration and operation of the device allow for better scientifically controlled experiments in
water distribution systems. Therefore, water quality data can be compared over time at
one location, among various locations in the same distribution system, and between water
systems. Figure 5-15 is a drawing of a PRS monitoring station.
M58 book.indb 117 11/17/2010 4:25:45 PM

M58 book.indb 118
Table 5-2 Summary of coupon protocols used for distribution system corrosion measures
Exposure
Protocol Name Coupon Geometry Mounting Arrangement Duration Comments References
ASTM flat coupon • Flat rectangular coupons Held on insulating 12 to 24 months • Generally used to monitor ASTM D2688-83,
standard stamped from sheet stock stem and inserted into corrosion processes Method B (1983a)
Corrosivity of water (0.5 in. × 4 in. [1.27 cm × flowstream of distribution on distribution system ASTM G1-81 (1981)
in the absence of 10.2 cm]). main at elbow or tee. materials. ASTM G46-76 (1976)
heat in transfer • Both sides exposed to • Used primarily with
water flow. mild steel or cast-iron
materials.
• Useful for pitting
evaluations.
Illinois State Water • Machined pipe nipples Fitted into a machined Minimum • Coupon fabrication ASTM D2688-83,
Survey (ISWS) (0.75 in. [1.9 cm] ID, PVC pipe section and 120 days requires extensive Method C (1983b)
Protocol machined 4 in. [10.2 cm] length). plumbed into bypass off recommended machining.
nipple test Exterior coated with distribution main. • Coupon processing is
epoxy resin (60 cm2 chemically based (acid
[9.2 in.2] exposed surface baths).
area).
Modified ISWS • Short pipe section Coupons gasketed and Coupons • Coupon processing is Reiber et al. (1988)
coupon sleeve (0.75 in. [1.9 cm] ID, 1.5 in. held in clear acrylic sleeve sacrificed at 30-, mechanically based.
tester [3.8 cm] length). (up to 10 coupons per 90-, and 180-day • Measurement precision
(University of • Exterior phenolic resin sleeve). Flow restricted intervals. improved by use
Washington) coated (20 cm2 [3 in.2] to coupon interior. Easy of multiple coupon
exposed surface area). coupon substitution. Flow exposures.

controlled to duplicate • Can be used to monitor
alternating turbulent and corrosion process on
stagnation conditions. residential plumbing
materials.
Water Research • Flat rectangular coupons Coupons mounted on Coupons • Developed for use with Williams et al. (1984)
Centre, Swindon, (75 mm × 12.5 mm × insulating stems and held sacrificed cast-iron coupons but
United Kingdom, 4 mm [3 in. × 0.5 in. × in flow control rig. at periodic suitable for use with other
coupon rig 0.15 in.]; 25 cm2 [3.8 in.2] intervals for up materials.
exposed surface). to 22 weeks. • Flow rig has high- and
low-velocity sections.
US Army Corps of • Flat rectangular coupons Mounted on insulating Not specified • Coupon surface exposure Temkar (1988)
Engineers Research similar to the ASTM stems in special pipe rig, not typical of pipe wall Temkar et al. (1987)
Lab tester D2688-83 standard. designed to hold a large hydraulics.
number of coupons under
controlled hydraulic
conditions.
Notes: ID = internal diameter.
PVC = polyvinyl chloride.
11/17/2010 4:25:45 PM
Drawing is courtesy of CRC Press (Cantor 2009).
Figure 5-15 Drawing of a PRS monitoring station

Photos courtesy of CRC Press (Cantor 2009).
Figure 5-16 Stacks of metal plates exposed to water replace pipe loops in the PRS monitoring
stations
These devices are similar to pipe loops; however, stacks of metal plates are held in
contact with the test water instead of actual pipes. A surface area of metal to volume of
water, similar to that in a pipe loop apparatus, is maintained (see Figure 5-16).
M58 book.indb 119 11/17/2010 4:25:47 PM

The metal plates are not to be confused with traditional coupons as previously
described. The plates in the PRS monitoring station are used for pipe film analysis after
adequate exposure to water, where chemical and microbiological films and scales can
develop on the metal surface. This type of analysis was described previously in this chap-
ter in the section entitled Examination of Pipe Scales. The metal plates are relatively easy
to place in the analytical instrumentation for analysis without damaging the scales.
These standardized monitoring stations are operated with a timer controlling the
water flow similar to the operation of pipe loops. Also, samples of water exposed to metal
after a stagnation period are taken for analysis similar to water samples taken from pipe
loops. Even with these similarities, PRS monitoring stations have the following advan-
tages over pipe loops (Cantor 2008):
• They are more compact than pipe loops and can, therefore, be placed more strate-
gically around the distribution system to capture the water quality at critical loca-
tions, such as at areas with high water age and at entry points.
• They can be set to use less water than suggested with pipe loop operations.
• Pipe film analyses can be performed in a relatively easy manner on the metal
plates—no pipes need to be excavated or cut open for analysis. The pipe film anal-
ysis gives information suggesting how older pipe scales in the water system have
formed and what chemicals and mechanisms are most likely at work in the par-
ticular water system.
• Because of their compactness and construction, the monitoring stations are more
accessible for use at all water utilities—large and small—and are more conducive
to routine water distribution system monitoring than pipe loops.
A detailed description of how to assemble and operate a PRS monitoring station
is given in the book Water Distribution System Monitoring: A Practical Approach for
Evaluating Drinking Water Quality (Cantor 2009).
Electrochemistry Loops__________________________________
Electrochemical (EC) monitoring for corrosion rates is a supplement to pipe loop and cou-
pon monitoring (metal release). Corrosion is an EC process, and electrochemistry can be a
powerful tool in its assessment. EC techniques can determine the underlying rate of corro-
sion as well as characterize the surface reactions that control or limit corrosion. In the past
decade, there have been substantial strides in hardware and technique development. EC
methodology has made the evolutionary adaptation from a purely laboratory-based technol-
ogy to an operational tool useful for compiling a corrosion history, screening a set of corro-
sion inhibitors, or optimizing a water quality regime for corrosion control.
The suitability of an EC methodology is dependent on the corrosion morphology; EC
corrosion measures may be inappropriate on surfaces subject to heavy pitting-type corro-
sion (i.e., mild steel or cast iron). But for uniformly corroding surfaces such as lead, copper,
solders, zinc, and brass, EC methods can often provide a nearly instantaneous measure
of the corrosion process. Unlike metal release measurements, which are cumulative, EC
techniques give a snapshot of the corrosion process. They define the corrosion rate at a
specific point in time and, hence, are of value in determining how short-term changes in
water quality and flow conditions influence corrosion processes or how a chemical addi-
tives program may be optimized to limit corrosion.
The EC methodologies that have found widest application in distribution system
corrosion assessment are those based on polarization measures. Descriptions of three
common polarization measures (potentiodynamic scans, linear polarization, and imped-
ance spectroscopy), along with application, limitations, and references, are summarized
in Table 5-3. A modified version of the potentiodynamic scan was used in the DC Water
corrosion assessment.
M58 book.indb 120 11/17/2010 4:25:47 PM

M58 book.indb 121
Table 5-3 Summary of electrochemical corrosion assessment methodologies
Operating Equipment
Name Data Form Principle Applications Precision Requirements References
Electrical resistance Resistance change Increased electri- On-line continuous Poor. Is useful only Slide wire probe Rohrback (1956)
in slide wire probe cal resistance on corrosion measure as a relative indica- Kelvin bridge Cooper (1986)
a corroding wire tor of change in Package cost:
is correlated to a corrosion rate <$5,000
decrease in cross-
sectional area due
to corrosion-
induced metal loss.
Potentiodynamic Anodic–cathodic Electrokinetic Point measure of Yields reliable re- Polarization cell Dean, France, and
scans current-potential interpretation of corrosion current sults on uniformly Potentiostat Ketcham (1971)
curve developed polarization data and Tafel con- corroding sur- Data logger Dean (1977)
from polarization — application of stants faces. Interpreta- Data interpretation
scans on both the Butler Volmer Pitting potential. tion of data from software
sides of the freely equation. pitting surfaces is Package cost:
corroding surface difficult. <$15,000
potential
Linear polarization Polarization resis- Single-point polar- Intermittent on-line Variable (dependent Polarization probe ASTM G59
tance (Pol Res). ization sequence— corrosion rate on accuracy of (two or three elec- Mansfeld (1977)
linear relationship measurements. Tafel constant) trode models)
between Pol Res Single-point poten-
and corrosion rate. tiostat
Package cost:
<$5,000
Electrochemical Impedance spec- Polarization fre- Corrosion rate Interpretation of Polarization cell Silverman (1986)
impedance spectros- trum (102–107 Hz) quency response Electrical model of model accuracy is Potentiostat Kending and Mans-
copy (EIS) of impressed al- corrosion surface controversial Frequency analyzer feld (1984)
ternating current Coating effective- Data interpretation EG&G (1989)
yields impedance ness software
spectrum Package cost:
>$25,000
Electrochemical Low-frequency Localized corrosion Remains a research Unproven Galvanic coupling Hladky and Dawson
noise changes in freely events emit a base tool cell (1981)
corroding surface signal character- Frequency analyzer Dawson and Uru-
potential istic of the magni- Zero resistance churtu (1986)
tude and type of ampmeter
corrosion process Package cost:
>$20,000
11/17/2010 4:25:48 PM
Table 5-4 Electrochemical corrosion data analysis software

Software
Package Application Comments References
CORFIT Capable of determining Tafel It is the original and remains the most Mansfeld (1973)
values and corrosion-current frequently used routine for polarization
densities from data obtained in data interpretation. Observed data is fit-
the non-Tafel region of typical ted to the electrokinetic model using a
polarization curves. least squares method. The usefulness of
the approach is restricted by the limited
number of data points the fitting routine
can handle. Program written in FOR-
TRAN.
Polcurr Similar application range as Uses an improved linearized least Gerchakov, Udey,
CORFIT. squares method of data fitting. Capable and Mansfeld
of handling up to 51 data points. De- (1981)
signed to run on VAX systems.
Polfit Capable of estimating anodic and Can provide point-by-point ohmic drop Shih and Mansfeld
cathodic Tafel slopes, corrosion- compensation (IR correction) for all ob- (1991)
current density, and Pol Res and served data points. Provides full graphic Jensen and Britz
of establishing the confidence representation functions. Capable of (1990)
limits for each of these param- simulation modeling of hypothetical
eters. polarization parameters. Written in
FORTRAN 4.1.
Baukamp Designed to aid the interpretation Utilizes systems approach. Limited to EIS EG&G Princ-
EIS soft- of EIS; capable of developing applications. eton Applied Re-
ware electrical model of corroding sur- search (1991)
face consisting of distinct electri-
cal building blocks.
Gamry soft- Capable of all standard DC polar- Software provides instrument control, Gamry Corpora-
ware pack- ization data interpretation; full data logging, and analytical functions tion (1992)
age graphics representation; real- for Gamry polarization instruments. Ca-
time data collection and interpre- pable of handling over 2,000 data points
tation features. simultaneously. Operates in Windows™
environment.
Recent electrochemical advances have been made in interpretive procedures of

polarization data. The utility of these procedures lies in their ability to characterize—with
a minimum of analyst effort—the form of the current-potential curve. Characterization
with minimal effort of current-potential curves is of significance in a drinking water con-
text because it allows for minimal perturbation of corroding surfaces that are sensitive
to even minor polarization events. Many of the recently developed software packages pro-
vide full graphic functions, sophisticated regression techniques for data fitting, and even
experimental simulation options based on modeled parameters. Some popular packages
are listed along with applications, limitations, and authors in Table 5-4.
EC assessment is flexible and can be adapted to a variety of experimental needs. In
many instances it has been possible to produce polarization cells that mimic the hydrau-
lics, geometry, and surface conditions of the actual corrosion surface. The object of a
polarization cell is to isolate a test specimen in such a way as to measure its response to
an electrical perturbation, while at the same time maintaining a contact environment that
closely replicates the freely corroding conditions of its in-service exposure. Figure 5-17
presents a photo of the polarization cell rig used in the DC Water corrosion assessment.
The EC monitoring option is generally simple to implement and may be rigged as part
of a larger pipe loop designed to study metal release. EC techniques may also serve an
M58 book.indb 122 11/17/2010 4:25:48 PM

individual loop designed solely to obtain corrosion rates. Corrosion rate data are usually
expressed as a penetration rate, MPY (mils per year of metal loss), or as an electrical cur-
rent exchange rate, µA/cm2. Either way, EC monitoring can provide a snapshot in time of
the corrosion process on a specific pipe specimen under a given water quality condition. It
is, effectively, the only technique that allows such a snapshot.
There is a great deal of flexibility in configuring pipe coupons used in EC monitor-
ing. Figure 5-18 shows a setup using a ¾-in.-diameter copper coupon with a total exposed
surface area of 20 cm2. Configured properly, an EC corrosion cell can accommodate
substantially larger pipe specimens. A principal advantage of EC monitoring is that pipe
specimens with existing corrosion scales can easily be accommodated in the testing pro-
cess. New pipes without corrosion scales are useful when doing a comparative assessment
of water quality effects on a particular metal type, but scaled surfaces generally give a
more accurate representation of corrosion on aged plumbing surfaces typical of in-service
conditions.
In Figure 5-18, the picture on the left is an actual setup of the schematic of EC polar-
ization on the right side. What makes this polarization cell unique is that it incorporates a
pipe segment as the working electrode. The Ag/AgCl reference electrode is used to track
the voltage changes on the working electrode, and the platinum counter electrode is used
to complete the circuit that impresses an electrical current on the working electrode. Water
flow rates through the cell can be adjusted to represent desired hydraulic conditions.
Conditioning of EC loops and test materials is important and may frequently extend
to 1 or 2 weeks, though it should be noted that conditioning of metal release loops often
exceeds 2 months.
EC corrosion rate monitoring requires some special planning. The essence of most
EC protocols is to impose a current on a specimen and measure the electrical potential
response of that surface. It is important to recognize that repeatedly imposing a current
on a corroding surface can actually change the EC nature of that surface, which in some
circumstances may limit the frequency with which measurements should be made. On a
fresh (new) plumbing surface, measurements can likely be made daily, even hourly, if the
underlying corrosion rate is relatively high. However, on aged surfaces where the underly-
ing corrosion rates may be exceptionally low, measurements may need to be restricted to
once daily, possibly every second or third day.
A variety of different EC protocols can be used to measure or estimate corrosion rates.
The most rigorous of these involves characterizing the Tafel behavior of the pipe specimen
by polarizing the test surface to sufficiently characterize the anodic and cathodic behavior
of the polarization curve (a polarization curve presents the surface potential of the test
specimen as a function of the impressed current in both a positive [anodic] and negative
[cathodic] direction). A Tafel assessment generally requires a high degree of polarization
and may generate the problems previously discussed. An alternative to the full polariza-
tion scheme required for Tafel assessment is a polarization resistance (Pol Res) test, which
requires a much smaller degree of polarization. Pol Res tests are conducted using a limited
polarization regimen that may be as little as ±5 mV around the freely corroding surface
potential. This relatively small degree of polarization does little to change the underlying
corrosion process. Typical procedures are to conduct Tafel tests once a week and on the
remaining days of the week conduct Pol Res tests based off the most recent Tafel data.
With the use of sophisticated computer software, provided by vendors such as
Gamry, data from Tafel and Pol Res tests can be quickly converted into corrosion rates.
Again, because of EC measurement variability, it is advisable to use several pipe specimen
replicates (recommend a minimum of three) to characterize the corrosion behavior of a
particular material under a specific water quality condition.
M58 book.indb 123 11/17/2010 4:25:48 PM

Figure 5-17 Electrochemistry circulation pipe loop setup
Chamber for
Reference
Electrode
Copper
Coupon Soldered wire
for
Potentiostat
Connection
Platinum
Counter
Electrode
Figure 5-18 Schematic of electrochemical polarization cell used for corrosion rate measurements
Figure 5-19 shows a typical polarization design. The setup contains four polarization
cells in which three of the cells contain copper coupons from the same pipe to provide
replicated values.
Figure 5-20 presents corrosion rate data obtained by an EC test rig using lead cou-
pons with lead (IV) scale from the DC Water system. Note the drop in corrosion rates once
orthophosphate was added to the water; however, lead concentrations took over 30 days
to show a substantial decrease.
M58 book.indb 124 11/17/2010 4:25:49 PM

20-L
Courtesy of HDR Engineering.
Figure 5-19 Typical polarization cell design
Echem Loop #1: Lead Coupon 1a

3.5 mg/L Chloramines
3.5 -4.0 mg/L 2.5 mg/L Orthophosphate
Free Chlorine
800
3.5 - 4.0 mg/L Chloramines 3.5 - 4.0 mg/L Chloramines
3.5 mg/L Orthophosphate
700
Corrosion Rate, mMPY
600
500
400 Coupon 1a
300
200
100
0
.
.
.
.
.
.
.
.
.
..
..
..
..
.
.
.
4/5/04 5/7/04 8/23/04 Date 1/30/06
Figure 5-20 Corrosion rates from lead pipe with lead (IV) scale
M58 book.indb 125 11/17/2010 4:25:50 PM

Premise Plumbing Profiles________________________________

In addition to pipe scale information, a utility may want to profile premise plumbing (in-
house plumbing and service line) using wet chemistry data taken from sequential samples
collected from the tap to the main. Premise plumbing profiles are useful to determine
metal leaching from the residential pipe scales, the stability of the pipe scales (i.e., how
readily the scale is breaking off the pipe wall), the concentration and types of particulates
disassociating from the scale, and microbiological activity with respect to nitrification
and other microbially influenced localized corrosion.
Premise profiles should be conducted at several homes in the distribution system in
order to establish a baseline. The process involves collecting pipe material data, which
includes pipe-construction material, pipe length, diameter, and location (i.e., in-house, ser-
vice line, or main). Once this information is gathered, the water is allowed to flow at an
inside tap for several minutes to ensure fresh water from the main is reaching the inside
sampling tap location. A sample is taken at the end-point tap, which will represent the
baseline information before stagnation.
After the flow, the water then sits in the line for 6 to 8 hr. At the end of the stagna-
tion period, water is collected sequentially in 1-L bottles at the home until the calculated
volume required to reach the main is met. Water should continue to run for several more
minutes to obtain current water quality conditions coming from the main. Approximately
200 mL from each 1-L sample are filtered in order to measure dissolved inorganic constit-
uents. This filtering will allow the user to obtain both particulate and dissolved inorganic
concentrations.
Finally, a water hammer is induced by repeatedly opening and closing the inside tap
at full velocity. Water is then collected from the tap from a representative sample inside
the home and in the service line (through volume calculations). This sample will then be
compared to the sequential samples just previously collected. Increased dissolved or par-
ticulate inorganic material such as lead, copper, and phosphate (if a phosphate inhibitor is
used), iron (galvanized material), and increased color may indicate an unstable scale for-
mation and the potential for these constituents to leach or break off and find their way to
the tap. Contrary to unstable scales, water hammer samples with low inorganics or similar
results as the sequential samples most likely indicate stable scales.
Inorganic-based profile samples are usually analyzed for metals relevant to the water
system, which may include lead, copper, iron, manganese, zinc, and aluminum. Other met-
als and inorganic anions (chlorides, sulfates, and so on) may also be collected depending
on the utility’s treatment or specific situation.
In addition to inorganic scales, organic or microbiological scales may be assessed.
Collecting sequential and water hammer samples for nitrites, AOC, total or free chlorine,
and HPC-R2As may provide the utility with information regarding microbiological inter-
actions that might increase localized corrosion. pH is also a recommended parameter to
measure for premise profiles.
A utility experiencing localized internal (inside the house) copper corrosion from
microbial activity may experience an increase in nitrite, drop in free ammonia (if adding
chloramine or systems with naturally occurring ammonia), a decrease in pH in low alka-
linity water, an increase in copper, and a decrease in disinfectant residual in a premise
plumbing profile after the 6-hr stagnation. Spikes in HPCs may occur during sampling;
however, the spikes may not necessarily indicate corrosion is occurring or that corrosion
is occurring at the particular location related to the sample number. For example, a spike
might occur at the third liter, but that spike may not necessarily mean the biofilm was
located at that point. A spike from this sample location indicates that biofilm exists; how-
ever, the biofilm could be from the third sample location to the first sample location.
M58 book.indb 126 11/17/2010 4:25:50 PM

Biofilm may not always detach from the pipe wall immediately. In addition, absence
of a spike in HPC does not always mean HPC is not present. The flow may be too low in
velocity to break up the biofilm and, therefore, HPC may never show up in the profile. Rely-
ing on the chemical parameters in conjunction with each other will probably provide more
accurate information. However, if a biofilm exists, it will most likely be dislodged during
the water hammer test.
Profiles in homes or buildings containing old galvanized lines where the zinc surface
coating has dissolved will show elevated iron levels. In addition, elevated lead levels have
been associated with this condition, although the mechanism for occurrence is not fully
understood. Figures 5-21 and 5-22 graph residential profiles conducted at homes.
Once the samples are collected, graphing the analysis of each constituent might be
helpful to quickly assimilate the data. Graphing total and dissolved metals is recommended.
Figure 5-21 is an example of a lead graph, with an overlay of total chlorine. Note how the
chlorine dissipated over the stagnation period in the service line. In this particular situa-
tion, the chlorine was reacting with the lead on the pipe surface to form lead oxide (PbO4).
Graphing the data helps visualize the profile. In Figure 5-21, it is obvious that the main
source of lead is from the service line in the form of a dissolved species, that is, lead (II).
Reservoir Profiles________________________________________
Changes in distribution water quality can occur from finished water reservoirs due to
long detention times, thermal stratification, or sediment buildup. The effects from these
changes may cause localized corrosion issues.
Long detention times are not uncommon in reservoirs as utilities try to provide
enough water for emergency purposes such as fire protection or a backup supply for treat-
ment and pumping malfunctions. In addition, reservoir short-circuiting creates pockets
or areas of extended water retention time. Extended retention time can be conducive to
bacterial growth as disinfectant residuals diminish. If long detention times are allowed to
continue, reduction in pH, alkalinity, DIC, dissolved oxygen, and ORP may occur in local
areas of the reservoir.
70 Lead 3.50
Service
60 e Main 3.00
In-House
Plumbing
Total Chlorine, ppm
50 2.50
Lead, ppb
40 2.00 Lead, ppb

Lead Filtered, ppb
30 1.50 Total Chlorine, ppm
20 1.00
10 0.50
0 0.00
r s lu s h
er
Li w
Li 2
Li 4
Li 6
Li r 8
Li r 10
Li 0
Li 12
L i 14
Li 17
L i 23
L i r 26
L i 29
Li 33
at Li r 4 3
m 3
r2
Ha r 6
ra
m
r
r
r
te
te
te
te
r
r
r
r
r
tD
F
te
te
te
te
te
te
te
te
te
er te
te
e
lin
Fi
se
Ba
Figure 5-21 Example of a lead profile at a residential home with a lead service line
M58 book.indb 127 11/17/2010 4:25:50 PM

35 700.0
Galvanize
In-House Lead
Main
Service Line
30 600.0
25 500.0
Total Iron, ppb

20 400.0
Lead, ppb
Lead Action Level (15 ppb)

15 300.0
10 200.0
5 100.0
0 0.0
0 1 2 3 4 5 6 7 8 9 10 12 15 18 22 X
Sample Volume, L
Lead, ppb Lead Filtered, ppb Total Iron, ppb
Figure 5-22 Example of a lead profile at a residential home with galvanized internal plumbing and
lead service line
It is important to have an understanding of the fluid dynamics in a reservoir or

storage tank to anticipate potential water quality changes that may occur. Conducting
reservoir profiles, especially during warm temperature months, will help determine the
extent of thermal stratification and the water quality in each zone as well as determine the
sediment buildup that could potentially affect Level 1 (baseline) corrosion water quality
parameters (see chapter 3).
Summary_________________________________________________
Utilities making changes to treatment or water sources should conduct preliminary tests
prior to making that change to ensure the treatment will be successful without any unin-
tended consequences. Bench testing is the easiest test to perform but will only provide
theoretical information related to water quality parameter changes; pipe loops or PRS moni-
toring stations will provide theoretical information related to scale formation and release.
Each of these tools used separately will provide the utility with some information
relating to theoretical water quality changes. However, conducting bench tests and pipe
loops followed by pilot testing (discussed in chapter 6) is the best way for ensuring the
true effectiveness of the treatment along with any unintended consequences. For further
information relating to monitoring treatment and source change impacts on corrosion
please refer to AWWA’s Managing Change to Avoid Unintended Consequences Related to
the Lead and Copper Rule Corrosion Control Practices (2005).
M58 book.indb 128 11/17/2010 4:25:50 PM

References________________________________________________
American Society for Testing and Materials Schock, M., and R. Giani. 2004. Oxidant/Disin-
(ASTM) D2688-83, Method B. 1983a. Stan- fectant Chemistry and Impacts on Lead
dard Test Methods for the Corrosivity of Corrosion. In Proc. of the AWWA Water
Water in the Absence of Heat Transfer Quality Technology Conference. Denver,
(Weight Loss Protocol). Philadelphia, Pa.: Colo.: AWWA.
ASTM. Temkar, P. 1988. Development of a Pipe Loop
ASTM D2688-83, Method C. 1983b. Standard System for Determining Effectiveness of
Test Methods for the Corrosivity of Corrosion Control Chemicals in Potable
Water in the Absence of Heat Transfer Water Systems. USA-CERL Technical
(Machined Nipple Test). Philadelphia, Report N-88/12. Champagne, Ill.: US Army
Pa.: ASTM. Corps of Engineers Research Laboratory.
ASTM G1-81. 1981. Recommended Practice for Temkar, P., R.J. Schloze, S.W. Maloney, and C.H.
Preparing, Cleaning and Evaluating Neff. 1987. Pipe Loop System for Evalu-
Corrosion Test Specimens. Philadelphia, ating Effects of Water Quality Control
Pa.: ASTM. Chemicals in Water Distribution Systems.
ASTM G46-76. 1976. Recommended Practice In Proc. of the AWWA Water Quality
for the Examination and Evaluation of Technology Conference. Denver, Colo.:
Pitting Type Corrosion. Philadelphia, AWWA.
Pa.: ASTM. Tesfai, F., P. Constant, S. Reiber, R. Giani, and
American Water Works Association (AWWA). M. Donnelly. 2006. Precipitate Forma-
2005. Managing Change to Avoid Unin- tion in the Distribution System Following
tended Consequences Related to the Lead Orthophosphate Addition. In Proc. of the
and Copper Rule Corrosion Control AWWA Water Quality Technology Confer-
Practices. Denver, Colo.: AWWA. ence. Denver, Colo.: AWWA.
Brodeur, T., and R. Flourence. 2007. Project Thomas, C., J. Kim, G. Korshin, J. Civardi, and
Seeks to Balance Corrosion Control, Scale R. Giani. 2004. Evaluation of Lead Leach-
Formation. Water World. November. ing Rates During Stagnation Using Real-
Cantor, A.F. 2008. Pipe Loop on Steroids! A New Time Corrosion Potential Monitoring and
Approach to Monitoring Distribution Sys- Modeling Methods. In Proc. of the AWWA
tem Water Quality. In Proc. of the AWWA Water Quality Technology Conference.
Water Quality Technology Conference. Denver, Colo.: AWWA.
Denver, Colo.: AWWA. Williams, S.M., R.G. Ainsworth, and A.F.
Cantor, A.F. 2009. Water Distribution System Elvidge. 1984. A Method of Assessing the
Monitoring: A Practical Approach for Corrosivity of Waters Towards Iron.
Evaluating Drinking Water Quality. Swindon, U.K.: Water Research Centre.
Boca Raton, Fla.: CRC Press.
Reiber, S.H., J.F. Ferguson, and M.M. Benjamin.
1988. An Improved Method for Corrosion
Rate Determination by Weight Loss. Jour.
AWWA, 79(2):71.
M58 book.indb 129 11/17/2010 4:25:50 PM

M58 book.indb 130 11/17/2010 4:25:50 PM

AWWA MANUAL M58
Chapter 6
Conducting Pilot
Studies and Monitoring
Effectiveness of Corrosion
Control Treatment
Richard Giani
DC Water
Introduction_____________________________________________
Once corrosion control treatment has been assessed and proved potentially successful,
a system is ready for a distribution system regimen. However, assessment tools like pipe
loops and jar testing may not have accounted for actual differences in true system design
and operations. If possible, a utility should conduct a pilot test, followed by full imple-
mentation with a startup and routine monitoring plan using tools discussed in previous
chapters.
This chapter will focus on the steps required to start up and maintain optimal water
quality parameters for corrosion control in the distribution system.
Table 6-1 lists the key points covered in this chapter.
131
M58 book.indb 131 11/17/2010 4:25:51 PM

• Pilot studies require more resources to operate but can provide true data related to interactions between
pipe scale and water quality; pilot studies take into account the distribution design, premise plumbing, and
unintended consequences.
• When conducting pilot studies, the utility’s priority is to ensure the protection of customers on the test system at
all times.
• Stability of pH minimizes corrosion. Entry-point water buffer intensities greater than 0.10 meq/L may be
necessary to maintain pH stability in the distribution system.
• A distribution system monitoring plan can include premise profiling, pipe loop or PRS monitoring station data,
electrochemistry data, pipe scale analyses, bench tests, and use of water quality models. The monitoring plan is
later streamlined after insight is gained from previous monitoring results.
• Action can be triggered if the system begins to drift from target water quality parameter levels.
• Water quality parameter target levels or ranges should be established for the distribution and entry point to
provide warning of deteriorating water quality. Target levels for water quality parameters should be more
stringent than those established by the regulating authority.
Conducting a Distribution System Pilot Study____________

Planning
The final step prior to implementing a full-scale treatment change is to conduct a distri-
bution pilot study. As previously mentioned, bench tests can assist with potential water
quality impacts related to the chemical and the water; pipe loops can provide data related
to direct reactions between the newly treated water and the existing (and new) pipe scales.
However, a distribution pilot study will also encompass information related to distribution
system configuration such as water velocities, existing biofilm along pipe walls, and water
usage. Basically, a pilot distribution study will provide true real-time data. However, con-
ducting a distribution pilot test takes time and proper planning to implement.
Pilot testing of the treatment scheme should be conducted on a small-plant scale and
in an isolated portion of the distribution system. Designs of isolated distribution systems
can vary from a few homes or a neighborhood that has been valved off from the rest of the
system and supplied by the pilot plant to a single pressure zone or a small section of the
distribution system that is fed only by a single entry point.
Use of bench testing and pipe-loop testing data prior to pilot implementation will be
useful in developing a plan to monitor treatment effectiveness while addressing potential
unintended consequences. It is important to capture as much information as possible dur-
ing the pilot study to assure the utility, the regulatory agency, and most of all the public that
the utility will benefit from the treatment and that at all times the public is protected.
It may take several weeks to plan and construct a pilot test before placing it in opera-
tion. During preparation, the utility will need to invest time to communicate, educate, and
receive buy-in from the customers served on the isolated system. In order to get custom-
ers’ cooperation, several public meetings may need to be held to assure customers that the
utility is ensuring their safety through vigorous monitoring and backup plans. It is also
important to recruit customers who are willing to volunteer to collect samples or who will
allow water quality technicians to enter their homes routinely to collect samples. Some
utilities offer volunteers incentives to participate such as credits to their water bills.
M58 book.indb 132 11/17/2010 4:25:51 PM

Conducting pilot studies and monitoring effectiveness of treatment 133
Selecting Operating and Monitoring Parameters

Pilot plants are designed to be self-sufficient and to operate with very little maintenance.
Most plants house automated chemical feed capabilities and online monitoring instrumen-
tation (e.g., to monitor pH, ORP, phosphate, chlorine, and so on). Grab samples may be nec-
essary to ensure chemical feed and online monitors are operating properly and to analyze
additional parameters (e.g., alkalinity, hardness, chloride, sulfate, and so on).
Changes in water quality from new treatment may weaken distribution system pipe
scales during the first several weeks to several months after the treatment change. There-
fore, chemical stability of water leaving the treatment plant is necessary (Figure 6-1) to
minimize rapid release of the pipe scale, especially by maintaining a steady pH.
Distribution monitoring. It is important to monitor several parameters within the
distribution system, for example, impacts of pipe scales, reduction in disinfectant levels,
and excessive precipitation.
Within the distribution system, particularly in older cast-iron mains, pipe scales are
extremely sensitive to changes in water chemistry. Monitoring for negative (and positive)
impacts of the distribution system scale related to treatment is extremely important. Rapid
scale release may cause increased discoloration in the water. In addition, scale release can
cause the micro-fauna protected within the scales to become dislodged, potentially caus-
ing temporary increases in HPCs. The number of positive total coliform samples may also
increase, as seen when the District of Columbia’s corrosion treatment changed from pH
adjustment to orthophosphate treatment (Figure 6-2). In August 2004, the DC Water began
adding orthophosphate to the distribution system. After the addition, a sharp increase
in positive total coliform samples followed. As scales began to stabilize, coliform levels
declined. It is important to anticipate potential consequences when conducting distribu-
tion pilot tests. In the case of Washington, D.C., the authority had developed an emer-
gency response plan (ERP) specific to the change in corrosion treatment (see case study
in Appendix B for details).
Reduction in disinfectant levels (especially chlorine) can coincide with scale release
as the chlorine will react with the newly exposed metals and bacteria on the remaining
pipe scale. In addition, if small amounts of ammonia are present, an increase in chlorine
odor may occur at the customer tap.
pH Fluctuations
Figure 6-1 Example of maintaining chemical (pH) stability at the entry point
M58 book.indb 133 11/17/2010 4:25:52 PM

10.0%
8.0%
Begin Addition of
6.8%
Orthophosphate
Percent Positive TC
6.0%
4.0% 3.8%
2.0% 1.7%
0.9%
0.5%
0.0% 0.0%
0.0%
Jun-04 Jul-04 Aug-04 Sep-04 Oct-04 Nov-04 Dec-04
Date
Figure 6-2 Percent positive total coliforms in DC Water in relation to the addition of orthophosphate
(June 2004 through December 2004)
Monitoring the distribution system for excessive precipitation is also important.

Precipitation could be related to pipe scale release, as the particulates will accumulate
at dead ends or low-velocity sectors. It can also be related to the formation of calcium
carbonate precipitation. If a phosphate inhibitor is added, orthophosphate can also form
precipitates (Figure 6-3). Phosphate levels in excess of 3.0 mg/L as PO4 can react with iron
oxide scales, calcium, fluoride, and aluminum (from aluminum-based coagulants) result-
ing in milky water (left bottle in Figure 6-3) (Tesfai et al. 2006).
Selecting and Preparing Sample Sites

Determining distribution sample locations is a major aspect of a sampling plan. Loca-
tions should be selected taking into account high-velocity and low-velocity areas, distance
from the treatment plant, and dead ends versus major intersections. The plan should also
include how the samples will be collected. Sampling locations may include reservoirs, fire
hydrants, residences, or strategically placed pipe loops and PRS monitoring stations.
Reservoir/clearwell monitoring. Samples collected from reservoir floors may pro-
vide useful information related to metals release (from iron-based mains) and precipitation
formation in the distribution system. Typically reservoir outlets are located approximately
1 ft from the floor of the tank. This placement provides an area below which sediment
buildup can be monitored and quantified.
Reservoir samples can be collected either by placing a sample tap 1 to 2 in. above the
reservoir floor or by using a depth sampler, which can be dropped down using a rope from
above (Figure 6-4).
M58 book.indb 134 11/17/2010 4:25:52 PM

Figure 6-3 Precipitated phosphate
Figure 6-4 Depth samplers used to collect grab samples from reservoirs
Fire hydrants. Sampling from fire hydrants is important for pilot testing because
such samples can provide a preliminary indication of potential precipitation buildup and
discoloration before degraded water quality shows up in residential tap samples. For this
reason, it is important to properly flush fire hydrants in preparation for sampling prior
to the start of the pilot project. If flushing crews are available, they should be placed
on standby to respond if rapid release of pipe scales creates water discoloration. Crews
should also be available to flush locations based on water quality data. For example, if
apparent color and iron levels begin to rise in a monitoring location, a crew should be dis-
patched to unidirectionally flush the area.
M58 book.indb 135 11/17/2010 4:25:54 PM

HPC comparison
Tap vs Hydrant
500
450
400
350
300
HPC, CFU
Hydrant
250
Tap
200
150
100
50
0
04
04
04
05
05
06
04
5
00
00
00
00
00
00
00
00
00
00
00
20
20
20
20
20
20
20
/2
/2
/2
/2
/2
/2
/2
/2
/2
/2
/2
9/
2/
0/
4/
0/
/
6/
28
/4
14
11
11
15
13
17
14
12
16
/2
/1
/1
/1
/1
/1
1/
10
12
9/
1/
2/
3/
4/
5/
6/
7/
8/
9/
10
11
12
10
11
Date
Figure 6-5 HPC results taken from a routine inside monitoring tap and a hydrant located next to
the property
Water quality results may also indicate a buildup of precipitates such as calcium
carbonate and aluminum phosphate in a low-velocity or dead-end area of the distribution
system. Flushing crews should be available to clean out the buildup before it reaches the
customers’ taps.
To utilize fire hydrants as sample locations, the hydrants must be checked a few
weeks prior to pilot testing to determine if they are operational. During this time, a 15-min
flush at a rate of 200 gpm will be necessary to remove any debris that has accumulated
in the hydrant service line and at the bottom of the hydrant chamber. This step should be
repeated two to three times during the next 2 weeks. Samples for water quality param-
eters, especially metals, should be collected and analyzed after the 15-min flush. Sam-
ple results should indicate very little precipitated metals. Common precipitates found in
unused hydrant service lines are iron, aluminum, calcium, and sometimes manganese.
Once a hydrant is placed into service for monitoring purposes, it is important to use that
hydrant at least once every 2 weeks to avoid water stagnation and accumulated metals.
When sampling, hydrants should be opened very slowly to release any trapped air in
the line. Once water begins to flow, slowly increase the velocity to 200 gpm for 5 min. The
time and rate of this velocity are not enough to scour the distribution line; however, the
goal is to dislodge any loose debris that could readily fall off of the pipe wall. After 5 min
of flushing, reduce the rate to approximately 10 to 25 gpm for sample collection. If testing
for total coliform, make sure to disinfect the hydrant coupling before flushing and use the
most probable number method or membrane filter method to provide a count of coliform
colonies. Using the presence–absence test does not provide a quantitative number and
that test may easily pick up a stray microbe from the hydrant chamber. HPC provides a
spectrum of bacteria that may be more representative of localized biofilm activity within
the pipe scales. Figure 6-5 illustrates HPC levels from an inside tap and a corresponding
hydrant located next to the premise. Notice the similarities in HPC levels. This specific
M58 book.indb 136 11/17/2010 4:25:54 PM

hydrant was flushed three times prior to initiating routine sampling. The first three hydrant
HPC samples were collected immediately after flushing the hydrant during the cleaning
process. Premise sampling was initiated when hydrant HPC levels stabilized and metals
analysis was typical of the rest of the distribution.
Hydrant flushing procedures should include the use of diffusers and disinfectant-
neutralizing chemicals if discharging into a storm drain or nearby stream. Make sure to
remove the diffuser (and neutralizer) when collecting the sample.
Residential sample taps. Chemical and biological reactions that occur within the
customers’ internal plumbing system may be entirely different from the reactions that
occur at the outside distribution mains. Pipe material inside a home is usually constructed
mostly of copper, with some homes containing galvanized or plastic pipes; fixtures may be
made of lead and brass; and service lines may also contain lead. Compared to the condi-
tions found in large iron-based mains, residential conditions such as warmer temperatures
inside a building or home coupled with long stagnation rates and higher pipe surface-
to-water ratios are more conducive to both microbial activity and metal reactions. As a
result, corrosion, chlorine uptake, and microbial levels may increase in residential sam-
ples. In addition, internal pipe scales will be much different compared to those in outside
mains. Therefore, it is important to capture all reactions beginning 2 weeks before and
throughout pilot testing.
Two levels of residential tap monitoring could occur during the pilot-testing phase:
daily/weekly routine tap sampling and monthly residential profiling. Residential tap sam-
pling procedures include removing the aerator to capture particulates in the water sample,
running the cold water for 30 sec at the rate of 1 to 3 gpm, and collecting samples for all
parameters described in the monitoring plan. Continue to let the water run for another 3 to
5 min, depending on the length of plumbing from the tap to the main, and collect another
set of samples. Once the residential tap sampling is completed, it is strongly suggested to
also collect samples at the closest fire hydrant or outside tap.
Residential profiling can be time-consuming and may be inconvenient for the cus-
tomer. Therefore, it is recommended to conduct a full metal profile at only one or two
locations prior to distribution pilot testing. To be more efficient, mini-profiles may be con-
ducted monthly once treatment has been initiated during the pilot phase. Mini-profiles are
advantageous because they only require the analysis of three to four samples and can be
collected by the homeowner.
When conducting a mini-profile, the utility should identify critical site-specific ser-
vice line and premise plumbing locations for sample analysis, such as first draw, inter-
nal plumbing, and the middle of the service line. Using pipe-volume formulas, calculate
the volume needed to capture these critical sample points. Mini-profiles begin by hav-
ing a customer flush the tap for 10 min, capture a sample, stagnate for 6 to 8 hr, and fill
up consecutive 1-L containers until the customer is pulling water from within the main.
Once completed, the customer should arrange for the utility to immediately pick up the
samples. Analysis should be conducted immediately from bottles that were determined
to be from critical locations. Critical samples should be filtered so particulate and dis-
solved metal concentrations are measured. The remaining noncritical bottles may be ana-
lyzed, held for future analysis pending critical sample results, or simply discarded. The
purpose of the noncritical sample bottles is to ensure that the homeowner collects the
water sequentially.
Mini-profiles should be compared to previous profiles in the home. Such a com-
parison will allow the utility to see changes that occur throughout the entire project.
Mini-profile homes should be chosen based on availability of volunteers, the volunteers’
locations within the distribution system (i.e., close to high/low flow areas, dead ends, and
so on), and the material of volunteers’ premise plumbing.
M58 book.indb 137 11/17/2010 4:25:54 PM

Pipe loops. If at all possible, pipe loops or PRS monitoring stations from previous
testing should be set up in the distribution system or at the pilot plant. Sampling should be
conducted at a minimum of 2 days per week for metals that would be expected in the pipe
scale. Parameters that are being measured in the distribution system during the project
should also be measured at the pipe loop or monitoring station influent.
Setting Water Quality Target Levels

Target levels or target ranges can be set for most water quality parameters that directly or
indirectly affect corrosion. When a target level is found to exceed a designated range or is
found to be outside of a designated range, that finding may instigate further investigation
or additional sampling. Several parameters outside of the target level or a single criti-
cal parameter outside the target level may instigate immediate action such as localized
flushing.
Target levels can be based on corrosion or microbial concerns or a combination of
both, and target levels can be different for inside taps and hydrants. If a parameter is
regulated (e.g., pH, alkalinity, and so on), target levels are usually established at a more
stringent level so as to ensure that regulated excursions are not violated. (Note that when
a daily average optimal corrosion control treatment water quality parameter is exceeded
at a specific site, the regulatory agency deems it an “excursion.”)
Examples of target parameters are found in Table 6-2. It is important to note that
target levels may vary based on the pilot-testing treatment, construction of distribution
pipe material, local regulations, and source water quality.
Developing an Emergency Response Plan

An ERP should be incorporated in the pilot-testing monitoring program. Major changes in
corrosion chemistry may cause scales to change, weaken (become soluble), or dislodge,
thus potentially causing an increase in metals and bacteria levels. A utility conducting
a pilot-testing project should review potential impacts and provide for rapid response
if they occur. For example, if a tap sample reveals a positive total coliform sample, the
homeowner should be made aware of the occurrence and educated ahead of time as to its
meaning. The utility rapid response to a positive coliform sample—following or in compli-
ance with the guidelines in the TCR—would be to ensure the sample is not E. coli or fecal
positive and to resample the location, along with upstream and downstream sites, within
24 hr. Elevated metals release may require the utility to temporarily provide sediment or
lead filters.
Record Keeping
During pilot testing, an enormous amount of data will need to be obtained in order to
determine if the pilot was successful and to optimize the treatment on a full-scale basis.
Water quality data, including water quality parameters, chemical dosages, profile infor-
mation, and so on, should be recorded in a database where the data can be filtered or
exported for graphical and statistical analysis.
In addition to water quality data, record keeping should include flushing records,
water quality complaints, emergency responses, customer notification records, and so on.
Recording complaints and flushing records can allow the utility to investigate and corre-
late these incidences with specific water quality parameters.
M58 book.indb 138 11/17/2010 4:25:54 PM

Table 6-2 Example of distribution target levels for water quality parameters routine monitoring
during pilot testing
Water Quality Parameter Target Levels
Aluminum, total (mg/L): Reduce phosphate precipitation ≤0.050 mg/L
(if applying orthophosphate).
Alkalinity (mg/L as CaCO3): Provides buffer to avoid pH Range (minimum based on adequate buffer
swing. intensity and maximum based on CCPP)
Ammonia (free NH3 -N mg/L): Chloramine systems only. 0.05 to 0.20 mg/L free NH3 -N
Low levels indicate nitrification; high levels indicate food
source for AOB.
Calcium precipitation (mg/L* as CaCO3): Avoids hydraulic ≤7 mg/L as CaCO3
stress if below target level.
Color, apparent (color units): Can indicate iron corrosion ≤25 color units
issues or phosphate precipitate above target level.
HPCs (CFU): Can indicate biofilm/microbial influence or ≤500 (hydrant) CFU
microbial release due to scale instability. ≤200 (inside tap) CFU§
Iron, total (mg/L): Can indicate iron corrosion above ≤0.30 mg/L
target level or indicate failure of sequestration chemical
if precipitate.
Precipitated manganese (mg/L):† Indicates failure of ≤0.05 mg/L
sequestration chemical (if applied) if above target level.
Nitrite (mg/L): Indicates nitrification, which can influence ≤0.050 mg/L
pH swing.
Orthophosphate (mg/L): Amount of reactive phosphate 1.0 to 3.0 mg/L
and precipitated phosphate. If high, could indicate
phosphate precipitation (if adding orthophosphate).
Dissolved orthophosphate (mg/L):† Amount of phosphate 1.0 to 3.0 mg/L
to react with lead.
pH: Used to maintain optimal passivation and pH stability. ±0.1 pH units at entry point
±0.3 pH units in distribution system
Chloride-to-sulfate ratio: Could increase lead solubility. ≤0.60 mg/L
Total coliform: Can indicate scale instability or increase Negative‡
in localized microbial activity.
Total chlorine (mg/L): A decrease may indicate microbial >1.0 mg/L
influence on pH or oxidation reaction with pipe scales.
* Total calcium precipitation = total calcium hardness – dissolved calcium hardness.
† Measured when the aluminum >0.050 mg/L, calcium precipitation >10 mg/L, iron >0.16 mg/L, color >100, or orthophosphate
>4.5 mg/L.
‡ If a total coliform sample is positive from inside tap, site will be notified close of business and instructed to flush inside plumbing.
Utility will conduct total coliform resample(s) the following business day.
§ If HPCs are >500 CFU in the tap sample, utility will advise the homeowner/occupant or responsible official to flush inside tap
regardless if any other target level is exceeded.
M58 book.indb 139 11/17/2010 4:25:54 PM

Monitoring the Effect of Corrosion Control

Treatment________________________________________________
Setting Optimal Water Quality Parameters

A corrosion control regimen for a water distribution system is typically determined from
the results of pilot studies. These results are used to set interim optimal water quality
parameters for corrosion control.
Entry-point and distribution fluctuations for water quality parameters that directly
affect uniform corrosion, such as pH and alkalinity, must be taken into account. A utility
may also want to include other parameters, such as those that indicate MIC, discoloration,
or particulate metals formation.
In most cases, because pH is the most sensitive parameter, the entry-point target
range should be ±0.1 pH units. Online monitors and SCADA systems can help maintain
entry-point pH stability. However, if these technologies are not available, manual pH mea-
surement will suffice.
The goal should be to keep historical records of pH, alkalinity, and temperature of
incoming source water. This data should, in turn, be correlated with chemical dosages for
adjusting the pH. Appendix A details these procedures for efficiently setting daily chemi-
cal dosages for pH adjustment and stability.
Adequate alkalinity is also important. Setting an alkalinity range can provide pH
stability in the distribution system. On the minimal end, alkalinity should be set to main-
tain a pH fluctuation of no more than ±0.3 pH units in the distribution system; and on the
maximum end of the range, alkalinity should be set to prevent precipitation of calcium
carbonate and avoid negative affects from elevated DIC.
When setting the minimal alkalinity range, a water’s natural buffering intensity
should be taken into account. Clement et al. (1998) found success in maintaining pH sta-
bility when the buffering intensity in Concord, N.H., was raised from 0.02 meq/L to 0.10
meg/L. DC Water has a large distribution system, maintains an average buffer intensity of
0.12 meq/L, and has successfully maintained a ±0.3 pH range.
Buffer intensity is dependent on pH and DIC. DIC is dependent on alkalinity. Fig-
ure 6-6 graphs the minimal entry-point alkalinity needed to maintain a buffer intensity of
0.10 meq/L at 20°C.
Alkalinity levels below 0.10 meq/L may cause pH decreases in the distribution system
and premise plumbing due to microbial activity. Kirmeyer et al. identified a decrease in
pH due to nitrification in poorly buffered waters (Kirmeyer et al. 2004). The city of Ottawa
(Ont., Canada) experienced elevated lead levels in an old section of the city. Further inves-
tigation identified elevated nitrite levels, higher HPCs, and diminishing pHs (Douglas and
Muylwyk 2006). It was thought that the nitrifying microbial community along the pipe wall
was causing localized depression of pH due to inadequate buffering (pH 8.7, alkalinity <25
mg/L as CaCO3, buffering intensity 0.05 meq/L). Ottawa decided to increase the pH to 9.2
(alkalinity ≈ 32 mg/L as CaCO3), which increased the buffering intensity to 0.1 meq/L. Con-
sequently, pH levels stabilized and lead levels dropped below regulated limits.
Developing an Initial Monitoring Plan

Once the distribution system corrosion control operations are initiated, aggressive moni-
toring should continue, similar to the pilot monitoring plan, until pipe scale stability is
accomplished. Stability normally takes several months but can take longer if the scale
continues to change. Intense monitoring should also be initiated to ensure the distribution
system has optimal water quality conditions systemwide, to ensure corrosion levels are
M58 book.indb 140 11/17/2010 4:25:54 PM

decreasing, and to determine that unintended consequences are not occurring. Chapter 3
provides insight as to the importance of specific water quality parameters and how they
should be utilized for corrosion monitoring.
Distribution monitoring locations are vital for corrosion assessment and should
include both inside residential taps and outside distribution taps (either hydrants or out-
side sample taps). The choice of locations should be based on several factors including
demographics (i.e., evenly distributed sample locations throughout the distribution sys-
tem), high-flow areas, low-flow areas, high velocity intersections, dead ends, locations
prone to complaints, and easy access to inside tap samples (e.g., businesses, schools, and
firehouses).
Samples should be collected at a frequency of once every 2 weeks, if possible. When
samples are being taken from both inside taps and distribution taps, collect samples
from the inside tap first. Samples collected from the distribution system, especially from
hydrants, potentially can cause scale disturbance, which might make its way into the
inside tap sample and thus provide false information.
Entry-point samples should include all parameters measured in the distribution sys-
tem for comparison and should be collected on a daily basis if possible.
Selecting a Sampling Technique

Premise profiling. Conducting routine profiles in customer homes is an effective
tool that can monitor scale changes, metals release, and microbial conditions in premise
plumbing. Inducing a weak water hammer in the premise plumbing by quickly opening
and closing a faucet several times after collecting profile samples is an effective tool to
determine the strength or stability of the scale at these locations. If the ratio of particulate
metals versus dissolved metals in the water hammer sample is greater than the ratio in
the profile samples, that finding would indicate that the scale is weak or still in transition.
Weaker scales can correlate to increased metals at the tap.
120
100
Alkalinity mg/L as CaCO3
80
60 Minimum Alkalinity
40
20
0
7 7.2 7.4 7.6 7.8 8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10
pH
Figure 6-6 Minimum alkalinity required to maintain a buffer intensity greater than or equal to
0.10 meq/L as a function of pH at 20ºC
M58 book.indb 141 11/17/2010 4:25:55 PM

3.5 mg/L free

chlorine
250
3.5–4.0 mg/L 3.5–4.0 mg/L Chloramines, 3.5 mg/L PO4
3.5–4.0 mg/L
Chloramines
Chloramines
200
Peak Lead Concentrations, ppb
150
Total Lead, ppb

Dissolved Lead, ppb
100
50
0
12/1/2003 3/10/2004 6/18/2004 9/26/2004 1/4/2005 4/14/2005 7/23/2005
Date of Lead Profile
Note: Water hammers can indicate stability of a pipe scale by the amount of particulate lead released during the process. Note the
decrease in the difference between particulate and dissolved lead after April 2005, indicating scales were stabilizing.
Source: Giani et al. 2005.
Figure 6-7 Lead profile water hammer graph taken from residential homes in the District of
Columbia from December 2003 to July 2005
Cl 2° HOCl ° OCl –
1.5
Water O
1
xidized
E, Volts
0.5 Cl –
Water R
-0.5 e duced
-1
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
pH
Source: Schock and Giani 2004.
Figure 6-8 Potential-pH diagram for 1 mg/L free chlorine, showing the speciation of the chlorine
system, the high ORP necessary for free chlorine stability, and the relationship to the water stability
boundary
M58 book.indb 142 11/17/2010 4:25:55 PM

As an example, Figure 6-7 shows particulate lead release during the water ham-
mer process in the internal home profiles in Washington, D.C. over a 2-year span, which
included chloramine, a temporary switch to chlorine, a return to chloramine, and finally
a switch to chloramine and orthophosphate (Giani et al. 2005). Once the orthophosphate
was added, the difference between the particulate lead and dissolved lead decreased dur-
ing the water hammers. In addition, overall lead levels also decreased. This trend would
tend to indicate that the lead phosphate scales strengthened. Also note how close the
particulate and dissolved lead levels were during the chlorine addition. During chlorine
addition, ORP levels were greater than 800 mV from elevated levels of chlorine (>3.0 mg/L
free chlorine). At a pH near 8.0, the water’s capability to oxidize metals (Figure 6-8) such
as lead (II) to lead (IV) was quite high (see DC Water Case Study, Appendix B, Figure B-4).
These elevated ORP levels appeared to stabilize the scale rather quickly. As soon as the
ORP decreased, lead levels began to increase and the difference between particulate and
dissolved lead also increased indicating an unstable scale.
Pipe scale analysis. Conducting XRD studies on internal piping scales quarterly
during the first several months of study may also prove useful in determining the types of
water chemistry changes that may be occurring. However, pipes must be harvested and
destroyed to perform this very informative test.
Pipe loops. If pipe loops or electrochemistry loops were set up during the pilot test-
ing, maintaining the loops during full system startup phases can provide useful informa-
tion with regard to scale sensitivity and metals release. These loops have been exposed
to the treatment several weeks ahead of the rest of the system and can provide advance
information as to the effectiveness of the treatment or unintended consequences during
that time frame.
Pipe loops can provide data indicating if the actual treatment plant is having effects
similar to those experienced in the pilot treatment process. For example, a pilot plant
using ferric chloride for coagulation may have a more effective mixing design than the
actual plant. Effective mixing usually leads to using a lower chemical dosage. The utility
may need to increase ferric chloride concentrations in the full-scale plant to achieve the
desired outcome. An increase of ferric chloride may cause an increase in the chloride-to-
sulfate ratio leaving the plant and therefore might have an impact that was not seen during
the pilot process.
PRS standardized monitoring stations. As discussed in chapter 5, PRS monitor-
ing stations are similar to pipe loops but have several advantages over pipe loops. The sta-
tions are convenient to use for all phases of monitoring, from monitoring the initial effects
of corrosion treatment through routine monitoring of distribution system water quality.
They combine the ability to monitor in a fashion similar to that of pipe loops with the abil-
ity to easily have pipe film and scale analysis performed for a more in-depth understanding
of chemical and microbiological mechanisms at work in the distribution system.
Using Water Quality Models

Use of water quality models such as the WATERPRO model and the Rothberg, Tamburini,
and Windsor (RTW, AWWA) model can assist utilities in predicting chemical concentra-
tions needed to hit target pHs. In addition, these models may also help utilities predict
CCPP in hard to slightly hard waters to avoid calcium precipitate formation if the target
pH is near the saturation point. Predicting CCPP is especially useful for surface water
utilities where source water pH, alkalinity, hardness, and temperature fluctuate on a daily
and seasonal basis. Groundwater systems with fluctuating pH and carbon dioxide levels
may also find the models useful.
M58 book.indb 143 11/17/2010 4:25:55 PM

Modeling should be verified by conducting CCPP bench tests and CCPP distribution
monitoring to determine true levels of calcium precipitation. (See chapter 5 for informa-
tion on how to conduct monitoring and bench testing for CCPP.)
Setting and Maintaining Target Levels

Water quality parameter target levels or ranges that are similar to target levels set in Table
6-2 should be established for the distribution and entry point to provide warning of deteri-
orating water quality. Target levels for water quality parameters should be more stringent
than those established by the regulating authority. If two parameters are outside target
ranges in a single sample, it may warrant further investigation or additional steps such as
localized flushing.
Single critical parameters may automatically warrant flushing or other action. Criti-
cal parameters may be directly related to corrosion treatment. Other parameters may not
be directly related but at certain levels outside of the target range may indicate a problem
is developing. For example, nitrite may have a typical target level of less than 0.02 mg/L
for systems adding chloramine disinfectant. However, concentrations that build up to
0.10 mg/L or greater may indicate significant nitrification and the concentrations may be a
sign that bacteria activity is high. In low-buffered systems, nitrite concentration buildups
may cause the pH to drop in the localized area of the microbial activity. Therefore, setting
a critical target level of nitrite greater than 0.10 mg/L as a single parameter could initiate
immediate flushing and quick remediation of the situation.
Target levels can also be used to initiate extra sampling of parameters not normally
measured as part of the routine distribution matrix. For example, a surface water system
that adds orthophosphate as a corrosion inhibitor has a reactive phosphate target range of
1 to 3 mg/L leaving the plant and in the distribution system. Aluminum was set at a typical
target level of less than 0.050 mg/L for both the entry point and the distribution system.
Aluminum levels exceeding the target level can indicate the occurrence and accumula-
tion of precipitated aluminum phosphate. Total phosphate readings may appear normal,
but in this situation measuring dissolved phosphate, or the phosphate that can still react
with lead and copper (i.e., not tied to the aluminum), becomes important to ensure enough
inhibitor is available for corrosion protection.
Refining the Monitoring Plan for Routine Monitoring

Once the distribution system corrosion control application is determined to be successful,
a routine monitoring plan should be established. This plan should address the continu-
ation of gathering baseline data, establishing target levels for water quality parameters
that may affect corrosion, and determining trends related to metals release and microbial
activity. The monitoring plan should be similar to the initial monitoring, but routine moni-
toring should be less frequent than initial monitoring.
Pipe loops or PRS monitoring stations should continue to operate on a daily basis,
but monitoring may be reduced to weekly or monthly frequencies, with plant and distribu-
tion action plans in place when target and excursion levels are exceeded. The frequency of
internal home profiles could also be reduced to quarterly or biannual profiles.
Triggering Action
Optimal corrosion control treatment water quality parameters, referred to as OCCTWQPs
in the LCR, are usually set by a system and approved by the regulatory agency with the
intent that corrosion control is optimized as long as the system maintains the specified
water quality parameter ranges. OCCTWQPs are parameters that have a direct impact on
corrosion and corrosion control treatment (e.g., pH, alkalinity, orthophosphate, and so
on). When a daily average OCCTWQP is exceeded at a specific site, the regulatory agency
M58 book.indb 144 11/17/2010 4:25:56 PM

will deem it an “excursion.” If two sites exceed the parameter on the same day, the agency
still considers it one excursion. Once an excursion occurs, daily monitoring must continue
at that site (or sites) until the daily average of the site-specific OCCTWQP is back to accept-
able levels. An additional excursion will occur each day that a daily average OCCTWQP is
exceeded. USEPA regulations allow up to nine excursions per 6-month monitoring period.
Therefore, utilities should have response plans to quickly bring parameters back into opti-
mal ranges. Figure 6-9 shows an example of a distribution excursion operations flowchart.
Target levels should be applied not only to the regulated OCCTWQPs but also to
additional parameters that have an indirect impact on corrosion treatment (e.g., nitrite,
chlorine, calcium precipitation, and so on).
In addition, target levels for OCCTWQPs should be more stringent than levels set by
regulators. Stricter target levels will allow the utility to react more efficiently to investi-
gate and improve the situation before excursion levels are reached. Some parameters can
have different target levels for residential taps versus distribution taps. Table 6-3 provides
an example of OCCTWQPs with regulated levels and OCCTWQPs /WQPs with target levels
that can be applied to both residential and distribution taps. When setting target levels, the
goal is to provide buffer to avoid an excursion of a regulated level.
Response by a utility is warranted when target levels are exceeded; however, the
response may not always need to be as immediate as the response would be when an excur-
sion of a regulated parameter occurs. For example, when an excursion level is exceeded,
resampling is initiated and followed by an immediate investigation of the situation. Flush-
ing a location until average OCCTWQPs are back to desired levels is likely, and it may take
several hours before readings return to normal.
Target level exceedances for nonregulated parameters may not need immediate
attention but may warrant localized unidirectional flushing of the mains within 24 to
72 hr. Customer taps that exceed target levels may require internal flushing of the lines
for 15 min over the next 1 to 5 days. When target levels temporarily are exceeded inside a
building but distribution samples taken from the main outside of the building are within
limits, the cause is most likely due to inadequate water usage inside the building. If inter-
nal target levels continue to rise, a more detailed investigation may be required to focus
on possible cross connections, microbial buildup due to improper plumbing disinfection
procedures, and so on. Figure 6-10 shows a flowchart taken from a utility that illustrates
standard operating procedures (SOPs) when target levels are exceeded for nonregulated
water quality parameters.
Table 6-3 Example of target and excursion levels for OCCTWQPs

OCCTWQP Regulated Levels Target Level
Average daily entry-point pH 7.5 ± 0.2 pH units 7.5 + 0.1 pH units
Individual-site average daily distribution pH >7.0 pH units >7.2 pH units
Average daily entry-point orthophosphate Range 0.5–3.0 mg/L Range 1.0–2.5 mg/L
Average daily entry-point alkalinity >20 mg/L Range 30–90 mg/L (minimum based
on buffer intensity of 0.10 meq/L and
maximum based on CCPP <7 mg/L as
CaCO3)
M58 book.indb 145 11/17/2010 4:25:56 PM

Regulated No No Further
further
OCCTWQP Exceeded action
Action
Yes
Flush Tap for 15 min
No Continue
Continue flushing
Flushingandand
OCCTWQP Exceeded Sampling
Sampling Until average
until Average
OCCTWQP Exceeded
is Below
is below regulated
Regulated
excursion level
Excursion Level
Yes
Hydrant Tap
Flush
Flush Local
local All Taps in Building
Hydrant in Direction from Entry to End
No No
Does Hydrant Exceed
Does hydrant exceed No Further
further Does Tap Exceed
Does tap exceed
OCCTWQP
OCCTWQP Regulated Levels
regulated levels action
Action OCCTWQP Regulated
Levels
Yes Yes
Investigate Distribution and Investigate Building

building
Treatment Plant (Cross-connection
(Cross connection investigation)
investigation)
Figure 6-9 Flowchart example of action taken when a regulated parameter (OCCTWQP) is exceeded
causing an excursion
M58 book.indb 146 11/17/2010 4:25:56 PM

No No Further
further
Target Range Exceeded
action
Action
Yes
Hydrant Tap
Flush Hydrant
hydrant Instruct Occupant
occupant
to Flush
flush tap
Tapfor
for15
15minutes
min
Resample 1 to 5 Days After Flushing
No No
OCCT Target
Target Level
Range Exceeded
Exceeded No
Further
Further Action
action
Yes
Hydrant Tap
Flush
Flush
Flush Local
local all
All Tapsininbuilding
taps Building
area
Area in Direction
in direction from
from entry
Entrytotoend
End
Resample Area Investigate Building

building
Within 1–5 Days
Figure 6-10 SOP flowchart for exceeding target levels
M58 book.indb 147 11/17/2010 4:25:56 PM

Summary_________________________________________________
Controlling corrosion can be very complex. It encompasses the production of a nonaggres-
sive water leaving the treatment plant. It also includes the maintenance of the integrity
of the water to mitigate potential negative impacts from bulk and localized chemical and
microbial reactions as the water makes its way through the distribution system to the
customer’s tap. As more stringent regulations are implemented, changes in treatment that
impact the chemistry of the water, in most cases will have some effect on corrosion sta-
bility. Having historical baseline data on key water quality parameters, having the ability
to conduct pilot testing in small portions of the distribution system or at the treatment
plant, and having the capability to monitor the distribution system using tools such as field
monitoring, profiling, and pipeloops helps a utility quickly optimize water quality while
avoiding unintended consequences.
References________________________________________________
Clement, J.A., W. Daily, H. Shorney, and A. Kirmeyer, G., K. Martel, G. Thompson, L. Rad-
Capuzzi. 1998. An Innovative Approach to der, W. Klement, M. LeChevallier, H. Bari-
Understanding and Improving Distribu- beau, and A. Flores. 2004. Optimizing
tion System Water Quality. In Proc. of the Chloramine Treatment. 2nd ed. Denver,
AWWA Water Quality Technology Confer- Colo.: AwwaRF and AWWA.
ence. Denver, Colo.: AWWA. Schock, M., and R. Giani. 2004. Oxidant/
Douglas, I., and Q. Muylwyk. 2006. Water Qual- Disinfectant Chemistry and Impacts on
ity Observations During Implementation: Lead Corrosion. In Proc. of the AWWA
A Case Study From the City of Ottawa. Water Quality Technology Conference.
Presentation. Denver, Colo.: AWWA.
Giani, R. 2006. Optimizing Corrosion Control Tesfai, F., P. Constant, S. Reiber, R. Giani, and
Treatment. New York: State of New York, M. Donnelly. 2006. Precipitate Forma-
Department of Health, New York Rural tion in the Distribution System Following
Water Association. Orthophosphate Addition. In Proc. of the
Giani, R., W. Keefer, and M. Donnelly. 2005. AWWA Water Quality Technology Confer-
Studying the Effectiveness and Stability ence. Denver, Colo.: AWWA.
of Orthophosphate on Washington D.C.’s
Lead Service Line Scales. In Proc. of the
AWWA Water Quality Technology Confer-
ence. Denver, Colo.: AWWA.
M58 book.indb 148 11/17/2010 4:25:56 PM

AWWA MANUAL M58
Appendix A
Achieving pH Stability
pH Stability_______________________________________________
Stability of pH involves tightening up the pH range at the entry point and providing enough
buffering capacity to avoid pH swings in the distribution system. Maintaining pH stability
minimizes metals release, dirty water complaints, taste-and-odor problems, and microbial
sloughing. For example, tightening the pH ranges will reduce the amount of iron released
from the pipe scale, thus reducing tuberculation and the formation of a protective barrier
for bacteria.
Maintaining a pH variation of +0.1 pH units from the target pH at the entry point is
the primary requirement for pH stability.
Alkalinity________________________________________________
In order to maintain a tight pH leaving the plant and a tight pH in the distribution system,
the alkalinity of the baseline water and finished water leaving the plant should have a
minimal alkalinity that causes the buffer intensity to be greater than 0.1 meq/L.
Sampling_________________________________________________
For pH stability, daily water quality parameter monitoring of pH and alkalinity is required.
Table A-1 summarizes the requirements for pH stability.
pH Stability Operations___________________________________
Maintaining pH stability can require a significant amount of operator attention, especially
in surface water treatment plants. However, operation can be conducted in a more concise
manner with the use of proper tools, jar testing, and creation of site-specific chemical dos-
age charts.
Startup Operations
Startup operations are probably the most difficult part of the corrosion treatment pro-
cess. The first few weeks normally involve numerous jar tests, monitoring, and creating a
149
M58 book.indb 149 11/17/2010 4:25:57 PM

chemical dosage chart or programming on-line supervisory control and data acquisition
(SCADA) systems.
For pH control, it is highly recommended to install an online SCADA system that can
maintain a pH leaving the plant. However, smaller utilities may not have the luxury of a
SCADA system and will have to adjust pH addition manually.
Table A-1 Summary of pH stability requirements

Requirement Comment
Entry-point pH +0.1 units from pH target
Example:
• Target = pH 7.5
• Range = 7.4 to 7.6
Distribution pH +0.3 units from pH target
Example:
• Entry-point target = 7.5
• Distribution pH range = 7.2 to 7.8
Alkalinity Range:
• Minimum alkalinity to maintain buffer intensity > 0.1 meq/L
• Maximum alkalinity to prevent CCPP > 7 mg/L as CaCO3
Sample frequency Daily
600
500
400
mL/min
Stroke 20%
300
40% Stroke
60% Stroke
80% Stroke
200 100% Stroke
100
0
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Speed, %
Figure A-1 Example of a dual-speed pump calibration curve
M58 book.indb 150 11/17/2010 4:25:57 PM

Appendix A 151
Pump Calibrations
The first task is to calibrate all corrosion treatment chemical feed pumps. It will be neces-
sary to know the pump curves for two reasons.
1. Determine efficiency and best settings of the pump.
2. Allow the operator to adjust pH adjustment chemical in an accurate and efficient
manner.
Figure A-1 illustrates a pump calibration curve for a liquid chemical feed dual-speed
pump.
Once pump calibrations have been completed, jar tests using the pH adjustment
chemical(s) should be conducted in order to determine the dosage needed to reach the
target pH. The target pH should have been predetermined in a desktop study.
Jar tests will provide the operator with an estimated chemical dosage (in mg/L)
needed to raise the pH to a certain level. Using the pump calibration chart, the operator
can determine the proper chemical feed setting needed to maintain the chemical dosage
that was obtained from the jar test. Fine-tuning may be required to reach the specific pH.
Once the target pH has been achieved, the operator will need to record the following
information on a water quality parameter (WQP) operations chart, a sample of which is
shown in Figure A-2.
• Raw pH
• Raw alkalinity
• Chemical dosage, in mg/L
Water Quality Parameter Operations Chart

The WQP operations chart is an extremely useful tool when adding a pH adjustment chem-
ical to the water. Basically, the chart organizes relationships between the incoming pH and
alkalinity water quality and the amount of chemical needed to reach the target pH. Such a
chart may take 2 months to 1 year (for surface water systems) to create and is site-specific.
No extra effort is needed to create a WQP chart. The data necessary to fill in this chart are
the same data that are required for monitoring corrosion control treatment. Figure A-2 is
a blank version of a WQP operations chart.
Setting Up a Chart
To begin filling out the WQP chart, place the chemical name and the system’s target pH in
the spaces provided near the top of the chart (refer to Figure A-3 throughout this discus-
sion). For example, in Figure A-3 the target pH is 7.8 and the chemical name is soda ash.
The next step is to set up the header row (pH) by taking the average raw water pH
(if groundwater) or average filter effluent pH (for surface water systems) and placing that
value in the first row, somewhere near the middle of the chart. In Figure A-3, the average
pH is 6.2.
Then, set up the header column (alkalinity) by taking the average baseline water
alkalinity and placing the value in the middle of the first column (in Figure A-3, the aver-
age baseline alkalinity is 25).
Now, go back to the raw pH value placed on the chart (in the middle of row 1). Start-
ing with the average raw pH value and moving from left to right in the header row, increase
the pH values by 0.1 units and write that value in the next column. Continue until the last
column is reached. Then, go back to the average pH (middle of row 1) and moving from
right to left, decrease the pH value by 0.1 units and write the value in the next column.
Continue until the first column is reached.
M58 book.indb 151 11/17/2010 4:25:57 PM

Repeat a similar procedure for the alkalinity column. However, unlike pH, there is
no 0.1-standard alkalinity increment to use in filling out the alkalinity values. Rather, the
alkalinity increment is dictated by average baseline alkalinity value. Table A-2 provides
the information on the increment, based on average baseline raw alkalinity values. In
Figure A-3, the alkalinity increment is +2.0 because the average alkalinity falls in the 20
to 40 range.
WQP Operations Chart for

pH/Alkalinity Adjustment
Chemical Name _______________ Target pH ________________

pH
ALK
Figure A-2 WQP operations chart for pH/alkalinity adjustment

Soda Ash
Chemical Name _________________ 7.8
Target pH __________________
pH
5.8 5.9 6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8
15
17
ALK
19
21
23
25
27
29
31
33
35
37
39
Figure A-3 WQP operations chart: soda ash and target pH of 7.8
M58 book.indb 152 11/17/2010 4:25:57 PM

Appendix A 153
Table A-2 Alkalinity increment (for use in WQP chart) as dictated by

average alkalinity baseline value
Average Alkalinity Value
(range in mg/L CaCO3) Increment for WQP Chart
1–10 +0.5
10–20 +1.0
20–40 +2.0
40–70 +3.0
70–100 +4.0
100–200 +5.0
200+ +10.0
Example of WQP Operations Chart Setup

To demonstrate how the WQP operations chart is set up, let’s run through an example. A
system is adding soda ash to raise the pH to a target of 7.8. The average raw water pH and
alkalinity are 6.2 and 25 mg/L, respectively. The pH increment for the chart will always be
0.1 pH units. Using Table A-2, we find that the alkalinity increment at 25 mg/L falls under
+2.0 increments. Now let’s fill in the first row with the pH values, and the first column
with the alkalinity increment values. At this stage, the chart should look like Figure A-3
discussed previously.
Now the chart is ready to be filled in. The best way to begin is to obtain previous data
from the past 2 months containing the baseline water pH and alkalinity, the entry-point
pH, and the dosage in mg/L of the pH adjustment chemical during each day.
From the historical data, we look for days when the entry-point pH readings were at
the target pH (i.e., 7.7 to 7.9). Each day that the target entry-point pH is reached, we obtain
the baseline pH and alkalinity for the same day.
Using the chart, we locate the baseline pH in the first row and run down the column
until we intersect the baseline alkalinity row. Then we can write down the chemical dos-
age (mg/L) in the box where the baseline pH and alkalinity intersect (Figure A-4). In this
example, let’s assume that the target entry-point pH was obtained using 35 mg/L of soda
ash and the baseline pH and alkalinity were 6.4 and 31 mg/L, respectively. The soda ash
dosage, 35 mg/L, is written in the box where the baseline pH and alkalinity intersect. Note
that these values should be written in pencil so changes can be made if necessary. In
summary, when the baseline pH at this treatment plant is 6.4 with a baseline alkalinity of
31 mg/L as CaCO3, a dosage of 35 mg/L of soda ash should hit the target pH of 7.8.
Now, we continue reviewing historical data until all information relating to days
when the entry-point pH was between 7.7 and 7.9. For each successful day, we place the
chemical dosage on the chart respective to the baseline pH and alkalinity that occurred
that same day. Note that if two or more days’ values fall in the same box, we write all val-
ues in the box and then take the average for all the values. That average value should then
be placed in the box.
The chart with the data written in is the starting point. Figure A-5 shows what a fin-
ished chart might look like.
Use of this chart on a daily basis will provide the operator with a quick chemical dos-
age reference for keeping the target pH within +0.1 units. All the operator needs to do is
collect baseline pH and alkalinity readings and an entry-point reading. If the entry-point
pH is outside the required range, then the operator uses the chart to find out the necessary
M58 book.indb 153 11/17/2010 4:25:57 PM


Soda Ash
Chemical Name _________________ 7.8
Target pH __________________
pH
5.8 5.9 6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8
15
17
ALK
19
21
23
25
27
29
31 35
33
35
37
39
Figure A-4 Filling in WQP chart with sample data

Soda Ash
Chemical Name _________________ 7.8
Target pH __________________
pH
5.8 5.9 6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8
15
17
ALK
19
21 35 29 26
23 33 30 29 30
25 34 32 30 30
27 34 33
29 41 37 35 34
31 40 35
33 44
35
37
39
Figure A-5 Working WQP operations chart
M58 book.indb 154 11/17/2010 4:25:58 PM

Appendix A 155
chemical dosage required to obtain the target pH. In most cases, the dosage will bring the
water within 0.1 pH units of the target pH.
During daily operations, if the operator comes across a box without a chemical
dosage, conduct a jar test to obtain that dosage and fill in the appropriate box.
Seasonal temperature variability of the baseline water can affect dosage rates. Oper-
ators of those systems that experience seasonal water temperature changes (i.e., surface
water systems) will create one chart for the winter, one for the summer, and one for fall
and spring.
Incorporating this information into a computer database can improve accuracy
and efficiency for obtaining target dosages. Once enough data is developed, the operator
can query multiple parameters for specific incoming pH, temperature, and alkalinity, and
the database will be able to identify average pH chemical dosages under those specific
conditions.
For example, if the incoming raw water of a plant had a pH of 6.2, alkalinity of 25,
and a temperature of 10°C, the operator could input these numbers into the database and
the computer could generate a report based on that system’s specific historical data and be
able to display that the average pH adjustment chemical concentration was 32 mg/L.
It is important to note that databases and chemical charts are only as accurate as
the data collected. Calibrating the pH meter and using good quality assurance and quality
control (QA/QC) practices are vital.
M58 book.indb 155 11/17/2010 4:25:58 PM

M58 book.indb 156 11/17/2010 4:25:58 PM

AWWA MANUAL M58
Appendix B
A Case Study:
Causes of and Actions
Taken to Control
Lead Release in the
Washington, D.C.
Distribution System
Laura Dufresne
The Cadmus Group Inc.
Richard Giani
DC Water
Introduction_____________________________________________
The increase in lead levels in drinking water in the District of Columbia (D.C.) represents
one of the most widely publicized corrosion problems to date. The D.C. problem focused
national attention on the importance of simultaneous compliance at a time when many
water systems were beginning to make treatment changes to comply with new drinking
water regulations. This case study was selected for two purposes.
157
M58 book.indb 157 11/17/2010 4:25:58 PM

158 Internal Corrosion Control in wATER Distribution Systems
1. To show how the increase in lead concentrations in D.C. drinking water was
related to a change in an atypical water quality parameter, oxidation-reduction
potential (ORP), as described in chapter 3.
2. To demonstrate how the utility, the regulating agency, and other local agencies
and experts involved utilized the tools described in chapter 5 to identify and
implement a successful corrosion treatment process and to highlight the utility’s
monitoring plan during the pilot study and full system implementation as dis-
cussed in chapter 6.
Water Treatment and Distribution in

Washington, D.C._________________________________________
When describing the D.C. drinking water system, it is useful to start with a list of prop-
erties that distinguish it from other water systems in the United States. First, the U.S.
Environmental Protection Agency (USEPA) has compliance oversight of the Safe Drinking
Water Act (SDWA) in D.C. through its regional headquarters in Philadelphia, Pa. Second,
the water distribution system, which is owned and operated by the D.C. Water and Sewer
Authority (DC Water), is a consecutive system. DC Water purchases treated water from
the US Army Corp of Engineers, Washington Aqueduct Division (Aqueduct). The Aque-
duct treats water from the Potomac River using mainly conventional processes. Although
buying and selling relationships among water systems are fairly common, the D.C. situa-
tion differs in that the Aqueduct also sells treated water to two utilities in Virginia (Falls
Church City and Arlington County) regulated by the Virginia Department of Health. Figure
B-1 shows the area served by the Washington Aqueduct.
The Aqueduct treats Potomac River water at two treatment facilities, the Dalecarlia
Water Treatment Plant and the McMillan Water Treatment Plant. Treatment techniques
include presedimentation, coagulation, flocculation, rapid-sand filtration, and primary
disinfection with chlorine. Alum is used for coagulation, and polyaluminum chloride is
added as a filtration aid. The Aqueduct uses lime to adjust the finished water pH at both
treatment plants. Since November 2000, ammonia has been added after the clearwells at
both plants to form chloramines for residual disinfection.
Figure B-1 The Washington Aqueduct service area
M58 book.indb 158 11/17/2010 4:26:00 PM

Appendix B 159
Table B-1 Key characteristics of the D.C. distribution system (based on 2005 data)
Characteristic Detail
Population served Approximately 570,000
Miles of pipe Approximately 1,300
Pipe material Cast iron (87%), ductile iron (8%), steel (2.5%), and prestressed concrete (2.5%)
Number of service connections Approximately 130,000
Service line material Predominantly copper, approximately 28% estimated to be made of lead*
Pressure zones 8
Storage Eight facilities owned by DC Water, three owned by the Washington Aqueduct.
Capacity ranges from 0.2 to 25 mil gal. Total storage capacity in the system
equals approx. 110 mil gal.
Pumping stations Four operated by DC Water, one operated by the Washington Aqueduct
Residual disinfectant Chloramines, with concentrations typically ranging from 3.0 to 3.5 ppm in the
distribution system
Corrosion control Orthophosphate, with concentrations typically ranging from 2 to 3 ppm in the
distribution system
pH 7.4–7.7
ORP 350–500 mV
*This estimate continues to decrease as DC Water implements their lead service line replacement program.
The D.C. water distribution system dates back more than 150 years, with pipelines
of varying age and composition. The majority of water mains are made of cast iron. Ser-
vice lines are predominately copper; however, DC Water estimates that approximately 15
percent are made of lead. Other key characteristics of the D.C. system are summarized in
Table B-1.
Chronology of Lead and Copper Rule Compliance

in D.C._ ___________________________________________________
Figure B-2 presents the 90th percentile lead levels in the D.C. distribution system since the
inception of the Lead and Copper Rule (LCR). The figure also shows important changes in
residual disinfection that impacted lead levels discussed later in this case study.
As shown in Figure B-2, D.C. drinking water exhibited moderate 90th percentile lead
concentrations ranging from 11 to 40 ppb in 1992 and 1993. A study, published in 1994,
recommended pH adjustment using lime for corrosion control. The 90th percentile lead
levels for the next two monitoring periods (January–June 1994, and July–December 1994)
were below the action level, allowing WASUA (Water & Sewer Utility Administration) (DC
Water’s predecessor agency) to begin reduced monitoring as specified in the LCR. Lead
levels remained below the action level through the summer of 2000.
Shortly following a switch from free chlorine to chloramines for residual disinfec-
tion in late 2000, lead concentrations at the customers’ taps increased sharply, with lev-
els reaching into the hundreds of ppb for some residents with lead service lines. USEPA
Region 3 brought in outside experts to review the optimal corrosion control treatment
(OCCT) designation of pH control using lime. They also worked with DC Water to conduct
additional monitoring to pinpoint the problem.
In January 2004, a Washington Post article brought national attention to the lead cor-
rosion problem in D.C. Shortly thereafter, the Technical Expert Working Group (TEWG)
was formed to coordinate and expedite research into the lead phenomenon. Members
M58 book.indb 159 11/17/2010 4:26:00 PM

80
70
90th Percentile Lead Level (ppb)
60
50
Beginning in November 2000,
WA switched from free
40 chlorine to chloramines for
residual disinfection
30
Beginning in August
20 2004, WA added
orthophosphate for
corrosion control.
10
0
Jan 92 - Jun 92
Jan 93 - Jun 93
Jan 94 - Jun 94
Jan 95 - Jun 95
Jan 04 - Jun 04
Jan 05 - Jun 05
Jul 92 - Dec 92
Jul 93 - Dec 93
Jul 94 - Dec 94
Jul 03 - Dec 03
Jul 04 - Dec 04
Jul 05 - Dec 05
Jul 95 - Jun 96
Jul 96 - Jun 97
Jul 97 - Jun 98
Jul 98 - Jun 99
Jul 99 - Jun 00
Jul 00 - Jun 01
Jul 01 - Jun 02
Jul 02 - Jun 03
Monitoring Period
Source: USEPA.
Figure B-2 History of LCR compliance in Washington, D.C.

of the group included representatives from USEPA Region 3, DC Water, the Washington
Aqueduct, Falls Church City, Arlington County, the Virginia Department of Health, the
D.C. Department of Health, and the Centers for Disease Control and Prevention (CDC). The
group also included technical experts from USEPA’s Office of Research and Development
Centers (ORD) and outside consultants.
Based on results of laboratory and pilot testing by TEWG members, orthophos-
phate was selected as the new corrosion control treatment. Full-scale application began
in August 2004 following successful pilot testing in a small area of the D.C. distribution
system. Within months of the new treatment, lead levels began to decline. DC Water met
the lead action level for the January 2005 to June 2005 monitoring period for the first time
since 2000 and has continued to meet the 90th percentile lead action level through the end
of 2008.
Using Corrosion Evaluation Tools to Determine the

Cause of Lead Leaching__________________________________
TEWG members used several corrosion evaluation tools to determine the underlying cause
of the lead increase in D.C. drinking water. Some techniques had been used extensively at
other utilities and were well documented. Others were developed or refined by TEWG mem-
bers specifically for the D.C. lead investigation. Evaluation techniques included extensive
monitoring at targeted residences, pipe loop studies, electrochemical testing, and pipe
scale analysis.
Detailed information on corrosion evaluation tools and preliminary findings for D.C.
were presented at a Sunday workshop during the 2004 AWWA Water Quality Technology
Conference (WQTC) in San Antonio, Texas, November 2004 (Schock and Giani 2004).
M58 book.indb 160 11/17/2010 4:26:00 PM

Appendix B 161
In July of 2005, USEPA published findings in a research newsletter (USEPA et al. 2005).
Readers are encouraged to refer to these publications for detailed descriptions of the cor-
rosion evaluation tools used by TEWG members to analyze the D.C. lead issue. A sum-
mary of key studies along with findings and implications are provided in the following
sections.
USEPA’s ORD evaluated the chemical composition of pipe scale material using X-ray
diffraction. Findings showed that the predominant scale material on lead service lines in
the D.C. distribution system was lead (IV) oxide, or PbO2. This key piece of information
helped researchers understand the reasons behind the lead release following the change
from free chlorine to chloramines.
A method for premise plumbing profiles was first presented at the 2004 AWWA WQTC
workshop (Giani and Edwards 2004). Premise plumbing profiles are created by stagnating
water for at least 6 hr at a residence, then collecting consecutive 1-L water samples from
the kitchen tap for analysis. Information on the plumbing system configuration is used
to determine where each sample was stagnating during the holding time. DC Water mea-
sured both total and dissolved lead of the water, creating detailed profiles from the water
main to the tap.
Premise profile results showed that the majority of lead leaching occurred in the lead
service line, and the majority of lead was in dissolved form as opposed to particulate form.
In early 2004, DC Water constructed six circulation pipe loops using lead service
lines extracted from the distribution system. Water was circulated through the loops for
a period of time and then stagnated for approximately 8 hr. After the stagnation period,
DC Water removed the water from the service lines and analyzed it for lead (dissolved and
total) along with a suite of other water quality parameters. Circulation loop testing was
done in accordance with methodologies from Internal Corrosion of Water Distribution
Systems (AwwaRF and TZW 1996). Lead levels measured during circulation loop testing
correlated well with premise plumbing profile samples from the lead service line.
In addition to using circulation pipe loops to collect baseline data, DC Water manip-
ulated various water quality parameters in the loops to evaluate their effects on lead
release. One important study (Schock and Giani 2004) showed that lead levels in pipe loop
samples dropped dramatically when DC Water changed from a chloramine residual to a
free chlorine residual. Premise profile results taken during the annual chlorine burn in the
spring of 2004 supported this finding. A comparison of peak lead levels in homes during
chloramination and during chlorination is shown in Figure B-3.
120
Peak Pb levels during chloramination
98 101
100 96
Peak Pb levels during chlorination
81
Dissolved Lead (ppb)
80
60 54
48
40
24 28
20
12 8
5 3
0
Resident 1 Resident 2 Resident 3 Resident 4 Resident 5 Resident 6
Figure B-3 Peak dissolved lead levels in homes during lead profiles
M58 book.indb 161 11/17/2010 4:26:00 PM

DC Water also constructed six electrochemical pipe loops to study the underlying
electrochemical reaction on the pipe surface and evaluate how reactions are affected by
water quality changes. They found that the electrochemical corrosion rate of lead pipe
specimens was very high, corresponding well with lead release rates in the circulation
pipe loops and in the premise plumbing profile samples.
Based on results of multiple evaluations, TEWG members postulated that the change
in ORP resulting from the conversion from free chlorine (>3.0 mg/L) to chloramines caused
the lead increase. ORP levels prior to the 2000 chloramine conversion averaged around
750 mV. Researchers believe that this relatively high level (along with the moderate pH
range of 7.4 to 7.8) was responsible for the development of an insoluble PbO2 scale on the
lead service lines. ORP levels after the conversion averaged around 450 mV. This reduc-
tion in ORP changed the form of the lead scale to a more soluble form, although the exact
reaction pathways are unknown. This theory is supported by research by Lytle and Schock
(2005), who demonstrated that PbO2 forms with time in chlorinated water (>3 mg/L) at pH
levels between 6.5 and 10.
Figure B-4 is a potential-pH diagram showing how the sequence of treatment changes
over the past decade formed and then destabilized the PbO2 passivating film. Point 1 repre-
sents the initial conditions of the early 1990s where there were some lead release problems.
An increase in free chlorine concentration in the mid-1990s moved the system chemistry
to approximately Point 2, causing formation of PbO2 (solid). The change to chloramines
for secondary disinfection in 2000 moved the ORP back into approximately the area of
Point 3, reversing the reaction.
Pb species = 0.015 mg/L; DIC = 18 mg C/L

I = 0; 25°C
Eh (volts vs SHE)
Figure B-4 EMF-pH diagram for Pb-H2O-CO2 system
M58 book.indb 162 11/17/2010 4:26:01 PM

Appendix B 163
Selecting the Best Treatment Technique__________________

The TEWG applied a wide variety of desktop, bench-scale, and pilot-scale techniques to
select the best treatment option for D.C. A summary of key findings from each study is
provided in the following sections.
The Washington Aqueduct and its expert consultants conducted a desktop study that
evaluated potential corrosion control options and identified those that had been applied
successfully to similar source water types. A complete copy of the 2004 study is avail-
able at http://www.epa.gov/dclead/CorrosionControl.pdf. Among the corrosion control
options considered were further increasing the pH and adding corrosion inhibitors such
as orthophosphate. The study found that further increase of pH using lime addition (the
existing treatment) would cause excessive precipitation of calcium carbonate in the distri-
bution system. Based on the wide use of orthophosphate and its success at similar water
systems, the report recommended orthophosphate treatment (by adding phosphoric acid)
to control lead corrosion. It also recommended a partial system test to evaluate the poten-
tial for adverse side effects of the treatment, such as red water, and a step-wise increase to
reach an optimal dosage of approximately 3 mg/L as PO4 (orthophosphate).
Both circulation and electrochemical pipe loop studies (as described in the previous
section) supported the desktop study recommendations. After a month of operation with
distribution system water, orthophosphate was added to the loops at varying concentra-
tions. Within a period of several days, the lead levels began to decrease for all loops with
orthophosphate. Figure B-5 displays results from DC Water’s circulation loop number 3,
showing how lead levels fell from more than 200 ppb to near 15 ppb after application of
orthophosphate treatment.
Results from Stagnation Loop 3

270
255 No treatment 10 mg/L Orthophosphate
240
225
210
195
180
Lead Concentration, ppb
165
150
135
120
105
90
75
60
45
30
15
0
9
4/5
2
9
6
5/3
0
7
4
1
6/7
4
1
8
7/5
2
9
6
8/2
8/9
6
3
0
9/6
/18
0
/4
/11
/25
/1
/8
/15
/22
/29
/13
/6
/20
/27
1/3
0
3/2
4/1
4/1
4/2
5/1
5/1
5/2
5/3
6/1
6/2
6/2
7/1
7/1
7/2
8/1
8/2
8/3
9/1
9/2
9/2
1/1
10
11
11
12
10
10
10
11
11
11
12
12
12
Date (2004)
Source: Rieber and Giani 2005.
Figure B-5 DC Water circulation loop testing results
M58 book.indb 163 11/17/2010 4:26:01 PM

Partial System Test of Orthophosphate__________________

Prior to applying the orthophosphate systemwide, a partial system test was conducted in
an isolated portion of DC Water’s distribution system. The test began in early June 2004,
with full-scale treatment scheduled to start in August 2004 pending test outcomes. The
objective was to identify any negative side effects of the orthophosphate treatment, such
as increased bacteriological activity or red water occurrences. Because passivation of
lead service lines was expected to take several months, researchers did not expect the test
to provide information on effectiveness of orthophosphate treatment.
The fourth high-pressure zone was selected as the test location because it encom-
passes a small area and a DC Water facility was available to install the temporary phos-
phoric acid feed equipment. Figure B-6 shows the location of the partial system test.
Two chemical feed pumps were installed inside the DC Water facility. Both were fed
from a feed tank located just outside the building. Online chemical monitors for pH and
ORP were installed before and after the injection point. See Figure B-7 for photographs of
the temporary chemical injection equipment.
Prior to the partial system test, DC Water staff took two important steps: (1) They
notified the public, and (2) they developed a detailed sampling and response plan. Several
weeks before the planned partial system test, DC Water mailed a fact sheet to all custom-
ers in the fourth high-pressure zone. DC Water also held two public meetings at a nearby
church on April 27 and 29 of 2004. At the meetings, DC Water, USEPA, and Washington
Aqueduct representatives described the partial system test and advised customers on what
to do in the event of red water (see Figure B-8 for handout distributed at that meeting).
Excerpts from DC Water’s partial system test sampling and response plan are shown
in Figure B-9. The plan was implemented in June and July of 2004. Results showed some
elevated heterotrophic plate counts at several sample sites and elevated color and iron in
about one-third of samples taken from fire hydrants. There were no customer complaints
of red water, however, in the test area. Members of the TEWG believed that these results
could reasonably be expected during the startup phase of phosphate-based corrosion con-
trol treatment and recommended that systemwide treatment begin as soon as possible.
Keeping a Watchful Eye on the System—Ongoing

Monitoring_ _____________________________________________
Since the orthophosphate treatment has stabilized, DC Water continues to implement vari-
ous monitoring programs that together provide a comprehensive picture of water quality
in the distribution system. Monitoring programs include the following:
• Total Coliform Rule (TCR) monitoring at inside taps or dedicated sampling
stations
• LCR monitoring at customers’ taps
• Stage 1 Disinfectants and Disinfection By-products Rule (DBPR) monitoring at
inside taps or dedicated sampling stations
• Routine monitoring at day cares and schools
• Water quality monitoring in response to customer complaints, repairs on new
mains, etc.
M58 book.indb 164 11/17/2010 4:26:01 PM

Appendix B 165
Figure B-6 Location of partial system test

Chemical Feed Pumps Chemical Feed Tank

Pump Controls Pretreatment and Posttreatment
Monitors for pH and ORP
Figure B-7 Photos of temporary phosphoric acid feed equipment
M58 book.indb 165 11/17/2010 4:26:04 PM

Fact Sheet: New Treatment to Address Lead in Water

Delivered to DC Water Customers in May 2004
On or about June 1, 2004 the Washington Aqueduct will add orthophosphate to the water supply. The chemical
will be added in two phases; first to the water supply for a small area in the Northwest quadrant, then later this
summer to the entire water distribution system. Orthophosphate will react with other minerals in the water to
create a protective lining on lead service line pipes, household pipes and plumbing fixtures, thereby reducing the
amount of lead that leaches from the pipes. It is possible that some residents may temporarily see rust-colored or
“red water” from their taps during this chemical change over.
What is red water?
“Red water” is reddish-orange water that occurs when iron dissolves from pipes into water delivered to the tap.
The presence of various forms of iron gives the water this color.
Why is my water this color?
Red water in your area is likely due to the addition of a chemical, orthophosphate, to the water to help decrease
lead levels. This color change is temporary and will go away with a few simple actions around the house and
with hydrant flushing by the DC Water.
Should I be concerned about drinking the water?
Iron is an essential nutrient and for most customers the water is safe to drink after a short period of flushing (see
below). EPA has a standard for iron in drinking water to protect against, for example, aesthetic (visual) and taste
effects of iron in drinking water. For customers with iron storage disorders, such as hemochromatosis, additional
iron in drinking water could present a health risk.
What should I do if I have red water?
If you experience red water...
Do not drink or cook with discolored water
Call DC Water’s 24-hour emergency services line at (202) 614-3400 and a flushing crew will respond
Run your cold water until it is clear
Do not use hot water - rust could settle in your hot water heater
Only wash laundry after your water runs clear
Flushing Instructions and Consumer Advisory
Homes with lead service line pipes should continue to flush their water until lead levels have decreased.
Homes with lead service lines
Draw water after 10 minutes of high water use activity (toilet flushing, showering, washing dishes or
clothes)
Then, flush tap for 60 seconds and collect and store water in the refrigerator for future use
Pregnant women, nursing mothers, and children under 6 years old should only drink filtered tap water
All homes
Flush tap for 60 seconds before drawing water
Use only cold water for drinking or cooking
Remove and clean the strainer/aerator/screen device on your faucet on a regular basis
Boiling water will not remove lead!
Use of water filters:
If you are advised to filter water, flush according to the instructions above before filling a water filtration
pitcher.
For more information… [contact names and numbers removed]
Figure B-8 Fact Sheet: New Treatment to Address Lead in Water
M58 book.indb 166 11/17/2010 4:26:04 PM

Appendix B 167
EXHIBIT 10 (PAGE 1 OF 4)
DC WATER
DEPARTMENT OF WATER SERVICES
DIVISION OF WATER QUALITY
1.1.1.1.1 Sampling Plan During Application of Corrosion Control Treatment

to the 4th High Service Area [Excerpts]
DC Water will perform the following distribution system monitoring during the application of corrosion
control treatment to the 4th High Service Area.
Distribution Water Quality Monitoring – During the phosphate injection stage, DC Water will moni-
tor from 15 fixed fire hydrant locations every week. Locations were chosen based on a hydraulic model in order
to obtain high and low flow areas throughout the pressure zone.
Distribution monitoring parameters are located on the sampling form in Appendix A of the sampling
plan. Hydrants will be flushed for approximately 3 to 5 minutes at a rate averaging 350 gpm. Hydrant flows will
be reduced to 25 to 50 gpm when the sample is collected. Total coliform, lead and copper samples will not be
collected from these sites.
The Washington Aqueduct will be collecting samples twice per week for total coliform at inside taps at
4 locations within 4th High (these are routine total coliform locations). In addition, pH, total & free chlorine and
orthophosphate readings will also be collected.
QA/QC – DC Water staff will use primarily Hach field testing equipment. Standard additions will be
conducted on all iron and phosphate samples, while remaining DC Water tests will be analyzed for recovery
using standards. pH meters will be calibrated daily to maintain slopes of –58 to –62.
Lead Profiling – Approximately 10 profiles have been conducted within 4th High under the influence
of chloramines. (These are supplemented by several other locations throughout the district outside 4th High.).
DC Water has also conducted approximately 7 lead profiles to study the effects of chlorine vs. chloramines.
Once the phosphate is injected at 4th High, DC Water will begin to conduct approximately two profiles
per week to determine the effectiveness of phosphate on lead and other metals. A listing of all test parameters is
located in Appendix B. It is expected that lead profiles will be conducted monthly at each home. When full sys-
tem treatment is implemented, we will continue to conduct profiles at these locations as well as other locations
district wide on homes that have been previously profiled.
In addition to the lead profile data, Water Quality Parameters listed in Appendix A will be collected at
the tap from each of the lead profile homes in the morning just prior to shutting down the water for lead profiling.
Residential Sampling – Tap Samples – Tap samples will be taken from at least 2 homes in 4th High
on a weekly basis. Samples will be monitored for all parameters listed in Appendix A. First draw, second draw
(determined by the calculated volume to reach the center of the lead service line) and a 10-minute flush will be
taken for each sample event. Residents will collect samples and DC Water will collect them on a weekly basis.
Flushing – DC Water has flushed 4th High in preparation to phosphate addition. Flushing was
Figure B-9 Excerpts from DC Water’s test sampling and response plan
(continued on next page)
M58 book.indb 167 11/17/2010 4:26:04 PM

EXHIBIT 10 (continued) (Page 2 of 4)
approximately 80% complete by June 1st. Flushing will continue throughout the outer limits of 4th High in June.
In order to determine effects that phosphate may have on the system (i.e., discolored water), flushing of the last
two zones will not occur until 2 weeks after phosphate addition. In addition to routine flushing, a special flushing
crew will be available to address discolored water complaints.
Data Evaluation – Data from 4th High and the bench scale testing will be evaluated by DC Water to
determine the potential effects phosphate treatment may have with regards to poor water quality (i.e., discolored
water) in the entire distribution system.
Water Quality Parameter Ranges – Interim water quality parameters were established as follows:
• pH: 7.7 + 0.3
• Orthophosphate: 1.0 – 5.0 mg/L
• Ammonia nitrogen monitor and report
• Nitrate/nitrite (N) monitor and report
DC Water will monitor the pH and orthophosphate leaving 4th High at least 6 times per day. In addition,
on-line monitors will continuously measure pH and orthophosphate as well as chlorine and turbidity.
Distribution monitoring results will be reviewed on a daily basis. Should interim water quality param-
eters fall outside the range at one distribution location (including total coliform and HPC results), it will be
considered an isolated incident. A flushing crew will be dispatched within 24 hours at the location and provide
extensive flushing within the area. Flushing events will continue every 12 hours until water quality parameters
fall within range.
If two or more distribution monitoring locations are outside of the interim WQP ranges (including total
coliform and HPCs), then sampling will be conducted at several other distribution sample locations, including
the Fort Reno Office to determine if a system wide incident may be occurring. If WQP’s at Fort Reno are outside
of the interim requirements, DC Water will contact the WA immediately in order to adjust the chemical treat-
ment. If treatment is adjusted and interim parameters (including total coliform and HPCs) are within range at
Fort Reno, but outside of the range at several distribution sites, extensive unidirectional flushing will commence
within 24 hours. Flushing will continue along with extensive monitoring until results indicate WQP’s are at
acceptable levels.
Positive Total Coliform Levels – If it is determined that a “system wide” incident is occurring in 4th
High with regards to elevated total coliform levels, DC Water shall take precautionary actions as stated in the
Total Coliform Rule. Upstream and downstream samples will be collected and analyzed for total coliform and
E-Coli. Monitoring at these sites will continue until all three sites are absent of total coliform and E-Coli. Exten-
sive unidirectional flushing will occur at and around all locations where a total coliform sample has indicated a
positive result.
Contact Information [Omitted]
Figure B-9 Excerpts from DC Water’s test sampling and response plan, continued
M58 book.indb 168 11/17/2010 4:26:04 PM

Appendix B 169
DC Water Department of Water Services Division of Water Quality

Lab Reporting Form
Sample LocatioN_ _____________________________________

Sample Type: (hydrant, residence, other)______________________________
(Circle one) Complaint or Routine
Date_ _____________________________ Time sample collected_ ___________ am/pm
Free Chlorine mg/L ____________ Copper __________ Standard ___________

Total Chlorine mg/L____________ 0.1 mL = 0.75 mg/L
pH ____________ slope ____________ +/– 0.05
Temp C____________ Nitrate NO3-N ____________
Alkalinity as CaCO3 ____________ Nitrites NO2-N ____________
Titrate to pH 4.5 Dissolved PO4 mg/L____________
Calcium Hardness as CaCO3____________ Standard____________
Color from red to blue. 0.1 mL = 0.5 mg/L
Add 50 mL deionized to sample +/– 0.05
Calcium Dissolved Hardness as CaCO3 Total PO4 mg/L____________
____________ Standard_____________
Color from red to blue. 0.1 mL = 0.5 mg/L
Add 50 mL deionized to sample +/– 0.05
(Filter) Comments: _______________________________
Iron mg/L____________ _________________________________________
Standard ____________ _________________________________________
_________________________________________
0.1 mL = 0.5 mg/L
+/– 0.05
Aluminum mg/L ____________
TDS mg/L ____________
ORP ____________
Sulfate mg/L ____________
Color ____________
HPC__________ Total Coliform__________
Ammonia NH3-N ____________
Lead mg/L ____________
M58 book.indb 169 11/17/2010 4:26:04 PM

EXHIBIT 10 (continued) (Page 4 of 4)

Appendix B.
Lead Profiling Parameters
Method Parameter
EPA 200.8 Aluminum

EPA 200.8 Antimony
EPA 200.8 Arsenic
EPA 200.8 Barium
EPA 200.8 Beryllium
EPA 200.8 Cadmium
EPA 200.8 Cobalt
EPA 200.8 Copper
EPA 200.8 Iron
EPA 200.8 Lead
EPA 200.8 Lithium
EPA 200.8 Manganese
EPA 200.8 Mercury
EPA 200.8 Molybdenum
EPA 200.8 Nickel
EPA 200.8 Selenium
EPA 200.8 Silver
EPA 200.8 Strontium
EPA 200.8 Thallium
EPA 200.8 Thorium
EPA 200.8 Uranium
EPA 200.8 Vanadium
EPA 200.8 Zinc
EPA 300.0 Bromide
EPA 300.0 Chloride
EPA 300.0 Fluoride
EPA 300.0 Nitrate
M58 book.indb 170 11/17/2010 4:26:04 PM

Appendix B 171
Table B-2 DC Water distribution system target levels (January 2006)

USEPA MCLs
Analyte (Primary or Secondary) DC Water Target Levels
pH 6.5–8.5 7.4 – 8.0
Alkalinity, mg/L as CaCO3 NA ≥30
Calcium precipitation, mg/L* NA ≤10
Total phosphate NA 0.5 – 4.5
Dissolved orthophosphate, mg/L† NA 0.5 – 4.5
Total chlorine, mg/L 4.0 2.0 – 4.0
Iron, mg/L ≤0.3 ≤0.16
Aluminum total, mg/L ≤0.8 ≤0.05
Manganese, mg/L ≤0.05 ≤0.04
Total dissolved solids, mg/L ≤500 ≤300
Sulfate ≤250 ≤100
Color, CFU ≤15 ≤15
Free ammonia, NH3 -N, mg/L NA 0.05 – 0.2
Nitrite, mg/L ≤1.0 ≤0.1
Total coliform‡ 5% total monthly samples Negative
HPC, CFU <500 <500 (hydrant)
<200 (inside tap)§
* Total calcium precipitation = total calcium hardness – dissolved calcium hardness.
† Measured when the aluminum >0.050 mg/L, calcium precipitation >10 mg/L; iron >0.16 mg/L; color >100; or total orthophos-
phate >4.5 mg/L.
‡ If a total coliform sample is positive from inside tap, site will be notified COB and instructed to flush inside plumbing. Utility will
conduct total coliform resample(s) the following business day.
§ If HPCs are >500 CFU in the tap sample, utility will advise the homeowner/occupant or responsible official to flush inside tap
regardless if any other target level is exceeded.
Collecting data is only the first part of the program. DC Water has established tar-
get levels for various water quality parameters that if exceeded, will trigger a specific
response. Table B-2 shows target levels current as of January 2006. The target levels are
always less than the maximum contaminant level (MCL) if there is one. Response flow-
charts (see example in chapter 4) detail actions to be taken to bring the water quality
parameter back under the target level.
References________________________________________________
American Water Works Association Research Lytle, D.A, and M.R. Schock. 2005. The For-
Foundation and DVGW-Technologiezen- mation of Pb(IV) Oxides in Chlorinated
trum Wasser (AwwaRF and TZW). 1996. Water. Presentation AWWA Water Quality
Internal Corrosion of Water Distribution Technology Conference. Denver, Colo.:
Systems. 2nd ed. Denver, Colo.: AwwaRF AWWA.
and AWWA. Rieber, S., and R. Giani. 2005. DC Water Lead
Giani, R., M. Edwards, C. Chung, and J. Wujek. Issues: Engineering Challenges, Engi-
2004. Use of Lead Profiles to Determine neering Solutions. Paper presented at the
Source of Action Level Exceedances from Water Environment Federation Technical
Residential Homes in Washington, D.C. Exhibition and Conference. Washington,
In Proc. of the AWWA Water Quality Tech- D.C., Oct. 31, 2005.
nical Conference. Denver, Colo.: AWWA.
M58 book.indb 171 11/17/2010 4:26:05 PM

Schock, M., and R. Giani. 2004. Oxidant/Disin- US Environmental Protection Agency (USEPA).
fection Chemistry and Impacts on Lead 2005. Research Newsletter: Determining
Corrosion. In Proc. of the AWWA Water the Causes of Elevated Lead Levels in D.C.
Quality Technology Conference. Denver, Washington, D.C.: USEPA.
Colo.: AWWA.
M58 book.indb 172 11/17/2010 4:26:05 PM

AWWA MANUAL M58
Appendix C
North American
Corrosion Control Needs
and Strategies:
A Summary of the 2008
AWWA-DSWQC Corrosion Survey
Melinda Friedman
Confluence Engineering, Seattle, Wash.
Anne Sandvig
Cadmus Group, Inc., Custer, S.D.
Quirien Muylwyk
CH2M HILL, Toronto, Ont.
ABSTRACT_ ________________________________________________
Implementation of the LCR in the United States has resulted in significant reductions in first
liter standing lead levels measured at the tap over the last 15 years. Proposed changes to
Health Canada’s Guidance for Corrosion Control, as well as provincial regulatory changes
related to lead in drinking water, are putting corrosion control on the agenda for utilities
in Canada. A North American Corrosion Control Needs and Strategies Survey, sponsored
by the AWWA Water Quality and Technology Division’s Distribution System Water Quality
Committee (DSWQC) and funded by the AWWA Technical and Education Council, was
conducted in the fall of 2008. This web-based survey summarized corrosion control objec-
tives, practices, effectiveness of practices (US utilities), and corrosion control needs (pri-
marily for Canadian utilities) for over 150 utilities in North America.
173
M58 book.indb 173 11/17/2010 4:26:05 PM

Survey Response Demographics and Characteristics_____

There were a total of 157 respondents who submitted surveys (120 from the US and 37 from
Canada). Of these 157 surveys, 119 were essentially complete (93 from the US and 26 from
Canada). Table C-1 lists a breakdown of the survey response.
Following is a summary of the 120 US and 37 Canadian utilities that responded to
the survey. The data is broken down by geography, population served, and general source
water quality ranges represented.
Geographic Breakdown
Figure C-1 lists the geographic breakdown by USEPA region for survey respondents in the
United States. The greatest number of US utilities that completed the survey was from
Region 5, followed by Regions 4 and 10. A comparison of the percentage of survey respon-
dents by USEPA region, versus community water systems (CWSs) in the United States
(based on the 2007 Safe Drinking Water Information System Federal Database [SDWIS-
FED]) is presented in Figure C-2. The greatest percentage of CWSs from the SDWISFED
data are located in Regions 4, 5, and 6, as compared to the greatest percentage of US sys-
tems that completed surveys, which were from Regions 4, 5, and 10.
Figure C-3 lists the geographic breakdown by province for Canadian survey respon-
dents. Ontario had the greatest number of systems (38 percent) represented in the survey
from Canada. No systems from Prince Edward Island (PEI), Yukon, or Nunavut completed
the survey. In Canada, Ontario and Quebec represent approximately 50 percent of the
country’s population.
Table C-1 Breakdown of survey responses

Complete Incomplete
Respondent Origin Surveys Surveys Duplicates Total
United States 93 20 7 120
Canada 26 11 0 37
Total 119 31 7 157
Number of Systems
Figure C-1 Geographic location of US participants by USEPA region (n=93)
M58 book.indb 174 11/17/2010 4:26:06 PM

Appendix C 175
35
30
Percent of Systems
25
20
15 SDWISFED Database
10 Survey Respondents
0
1 2 3 4 5 6 7 8 9 10
EPA Region
Figure C-2 Percentage of CWSs in SDWISFED (n=52110) versus percentage of survey respondents
(n=93) by USEPA region
Number of Systems
Figure C-3 Geographic location of participants by Canadian province/territory (n=26)
Population Breakdown
The population breakdown of the survey indicated that (Figure C-4):
• Systems from all population categories completed survey responses for both the
US and Canada.
• For both the US and Canada, the largest number of systems that completed sur-
veys were in the greater than 100,000–500,000 population category (33 systems in
the US, 6 systems in Canada).
General Source Water Quality Ranges

The Corrosion Survey requested information on source water characteristics for the two
largest sources of supply for each system. Figures C-5 through C-7 display source water
characteristics for Source #1 for pH, alkalinity, and DIC (reported). In summary:
M58 book.indb 175 11/17/2010 4:26:06 PM

Number of Systems
Figure C-4 Number of systems by population category (n=93 [US]; n=26 [Canada])
Number of Systems
Figure C-5 Range of source water pH (n=88 [US]; n=25 [Canada])
• There is utility representation in all pH ranges (with the exception of higher pH

source water greater than 9.5).
• The greatest number of systems in the US reported source water pH less than 7.2.
• In Canada, the greatest number of systems reported source water pH in the range
of greater than 8.0 to 8.5.
• A smaller subset of systems reported DIC levels. Further evaluation of the data
will include estimates of DIC based on pH and alkalinity for those systems that did
not report a DIC value.
M58 book.indb 176 11/17/2010 4:26:07 PM

Appendix C 177
Number of Systems
Alkalinity Range, mg/L as CaCO3
Figure C-6 Range of source water alkalinity (n= 85 [US]; n=21 [Canada])
18
16
16
14
Systems
12
Systems
10
8
# of of
6 US
Number
6 5
Canada
4 3
2
2 1 1 1
0
<=5 >5-10 >10-25 >25
Dissolved Inorganic Carbonate Range, mg/L CO3
Figure C-7 Range of source water DIC (n= 28 [US]; n= 7 [Canada])
CORROSION CONTROL OBJECTIVES_ __________________________

The objectives of corrosion control were summarized and evaluated by first segregating
utility responses into those utilities that reported they currently utilize corrosion control
and those utilities which do not. Figure C-8 displays this breakdown for Source #1. As
would be expected, the majority of US systems practice corrosion control (due to imple-
mentation of the LCR), and for Canadian systems, the majority do not practice corrosion
control. However, a relatively high percentage of Canadian systems that completed the
survey did practice corrosion control (38 percent of Canadian survey respondents).
M58 book.indb 177 11/17/2010 4:26:08 PM

Number of Participants
Figure C-8 Number of systems practicing and not practicing corrosion control for Source #1 (n=93
[US]; n=26 [Canada])
Table C-2 Typical source water quality parameters—utilities considering corrosion control
Source
Source Water
Water Alkalinity SW or
pH Range Range Utility ID Population GW Objective
US—Utility 1 >7.6–8.0 >50–100 687154940 >10,000–50,000 SW Lead
Can—Utility 2 >8.0–8.5 >100–200 692975061 >500,000 SW Lead
Can—Utility 3 >7.6–8.0 >50–100 693853177 >10,000–50,000 SW Lead
Can—Utility 4 >8.0–8.5 >200 693865240 >10,000–50,000 GW Lead
Can—Utility 5 >8.0–8.5 >100–200 694514034 >10,000–50,000 SW Iron and Manganese
Systems that Do Not Practice Corrosion Control

For those utilities that indicated they did not currently implement corrosion control,
there was a subset of systems that was considering implementation. In the US, where the
LCR has been implemented for several years, only 1 system of 20 that currently do not
practice corrosion control was considering implementing it, with the objective of reduc-
ing lead levels. In Canada, 4 systems of 16 that reported they do not practice corrosion
control said they were considering initiating it. A summary of source water quality pa-
rameters for these systems and their objective for initiating corrosion control are listed
in Table C-2.
Systems that Use Corrosion Control

For those systems that currently use corrosion control (74 utilities in the US and 10 utili-
ties in Canada), in general:
• For US respondents, control of lead and copper levels was the primary purpose
cited. Other reasons listed included iron and chloramines stability (with lead as
primary purpose), maintenance of water mains, compliance with regulations, con-
trol of TTHMs by having lower pH, mainline degradation, infrastructure protection,
M58 book.indb 178 11/17/2010 4:26:08 PM

Appendix C 179
Table C-3 Purpose for corrosion control at Canadian utilities with a corrosion control program
Corrosion Control Process
Province Purpose for Corrosion Control (for Source #1)
British Columbia Copper pH adjustment
Manitoba Iron, manganese, acidity pH adjustment
Manitoba Lead Phosphates
New Brunswick Iron Sequestering agent (SeaQuest)
Nova Scotia Iron, manganese Phosphate, pH and DIC adjustment
Nova Scotia Lead, copper pH adjustment
Ontario Iron, manganese Silicate
Ontario Lead, copper pH adjustment
Quebec Lead, copper, iron, manganese Silicate, pH adjustment
Saskatchewan Positive LSI Lime softening
and a lead hazard reduction program. These answers indicate that since imple-
mentation of the LCR, there is more focus on simultaneous compliance and infra-
structure maintenance goals that can be achieved in association with control of
lead and/or copper in drinking water.
• For Canadian respondents, other reasons cited included: Langlier saturation
index (LSI), acidity, microbiological quality, and chlorine residual. Based on these
responses, the primary focus for these systems in Canada is currently the LSI and
calcium carbonate stability versus control of corrosion and solubility of metals.
Table C-3 lists the 10 Canadian utilities that indicated they practice corrosion con-
trol, with the purpose for their program identified. Utilities from several provinces are rep-
resented, with more systems controlling for iron and manganese in the system than lead.
Corrosion Control Practices____________________________

Those systems that implement corrosion control treatment were asked to identify their
treatment method. The corrosion control treatment method for Source #1 is displayed in
Figure C-9.
• In the United States, pH adjustment (using chemical feed systems) was practiced
at the greatest number of utilities, followed by phosphate addition. Three utilities
in the US listed aeration in addition to pH adjustment.
• In Canada, pH adjustment had the greatest number of respondents.
• In many cases, more than one corrosion control method was listed (pH adjustment
and phosphate).
• One of the initial survey questions asked whether corrosion control was practiced
or not (Question 5). Results from this question indicated that 19 US and 16 Cana-
dian utilities did NOT practice corrosion control. However, another survey ques-
tion (Question 13) asked what corrosion control treatment was practiced, with “no
corrosion control treatment” listed as one of the options. In this later question,
only 10 US and 7 Canadian utilities listed that no corrosion control treatment was
practiced for Source #1. The difference may be due to systems that indicated they
use corrosion control treatment processes (pH adjustment, aeration, etc.) but may
actually use these processes for purposes other than corrosion control.
M58 book.indb 179 11/17/2010 4:26:08 PM

Number of Systems
Figure C-9 Corrosion control treatment used for Source #1
Corrosion Control Effectiveness and Needs_____________

An assessment of the effectiveness of US corrosion control practices (using comparisons
of 90th percentile lead and copper levels as a function of the type of corrosion control
treatment), and corrosion control needs for Canadian systems are presented in the follow-
ing section.
US Utilities
Overview of 90th percentile lead and copper levels. Figure C-10 and Figure C-11
show the percentile distribution plots of the 90th percentile lead and copper levels, respec-
tively, for US utilities that responded to the survey. The median 90th percentile lead level
for the entire data set was 0.003 mg/L, and the median 90th percentile copper level was
0.16 mg/L. No systems that participated in the survey exceeded either the lead or copper
action levels (AL). The median 90th percentile lead level is five times lower than the lead
AL, and median 90th percentile copper level is eight times lower than the copper AL, sug-
gesting that corrosion control treatment is effective for both lead and copper.
100
90
Percen�le Distribu�on
80
70
60
50
40
30
20
10
0
0.000 0.003 0.006 0.009 0.012 0.015
90th % Pb, mg/L
Figure C-10 Percentile distribution plot of US 90th percentile lead levels (n=85)
M58 book.indb 180 11/17/2010 4:26:09 PM

Appendix C 181
100
90
Percenle Distribuon
80
70
60
50
40
30
20
10
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
90th % Cu, mg/L
Figure C-11 Percentile distribution plot of US 90th percentile copper levels (n=84)
Effectiveness of treatment approach. Corrosion control effectiveness for US util-

ities was assessed by comparing 90th percentile lead and copper levels as a function of
finished water pH range (for systems that identified pH adjustment as their corrosion con-
trol treatment approach—this group does not include any systems that use phosphates or
silicates), and by evaluating 90th percentile lead and copper levels for systems that identi-
fied phosphate use as their corrosion control treatment approach (some of these systems
also practice pH adjustment as part of their phosphate treatment program). It is important
to note that this evaluation is overly simplistic because it does not yet account for major
sources of variability within the data set such as those systems with and without lead
service lines, source water type, population served, etc. Aggregation of the data set and
more detailed analyses will be conducted in the future. The current analysis shown in Fig-
ure C-12 through Figure C-17 is based on treatment data for Source #1 only.
Figure C-12—90th percentile lead levels as a function of pH
(phosphate systems not included)
• Thirty-three systems that provided 90th percentile lead levels indicated that they
use pH adjustment (no phosphates or silicates) for corrosion control.
• Systems with finished water pH in the range of greater than 8.0 to 8.5 had the low-
est 90th percentile lead levels.
• Systems with finished water pH in the range of greater than 7.6 to 8.0 had per-
formed better than systems with pH in the range of greater than 7.2 to 7.6.
• It is unclear why systems with finished water in the pH range greater than 8.5 to
9.0 had higher 90th percentile lead levels compared to systems in the pH ranges of
greater than 7.6 to 8.0 and greater than 8.0 to 8.5. Further analysis of the data set
indicated that 29 percent of the systems within this subset (finished water in the
pH range greater than 8.5 to 9.0) had lead service lines. Comparatively, 20 percent
of systems reported having lead service lines for the entire data set. Additional
potential contributing factors will be evaluated in the future.
M58 book.indb 181 11/17/2010 4:26:10 PM

100
Percenle Distribuon
75 pH>7.2-7.6
pH>7.6-8.0
50
pH >8.0-8.5
25 pH>8.5-9.0
0
0 0.003 0.006 0.009 0.012 0.015
90th Percenle Pb Levels, mg/L
Figure C-12 90th percentile lead levels as a function of pH (n=33) (phosphate systems not
included)
Figure C-13 and C-14—90th percentile lead levels for systems using
phosphates
• Forty-two systems indicated that they add phosphates for corrosion control. How-
ever, only 40 of these systems provided 90th percentile lead data, and therefore
only 40 systems are included in the analysis.
• Many of these systems also practice pH adjustment (effectiveness of phosphates
as a function of finished water pH to be evaluated later).
• The median 90th percentile lead level for systems that add phosphates is
0.002 mg/L.
• Figure C-14 provides a comparison of 90th percentile lead levels as a function of
type of phosphate used. The data suggests that all products performed similarly
between the 50th and 80th percentiles of the data set, but zinc orthophosphate
use resulted in lower 90th percentile lead levels above the 80th percentile of the
data set.
100
Percenle Distribuon
75
50
25
0
0 0.003 0.006 0.009 0.012 0.015
Figure C-13 90th percentile lead levels for systems using phosphates (n=40)
M58 book.indb 182 11/17/2010 4:26:11 PM

Appendix C 183
100
90
Ortho/poly blend
80
Percenle Distribuon
70 Phosphoric acid
60
50 Polyphosphate
40
30 Zinc Ortho
20
Proprietary
10 Orthophosphate
0
Zinc Orthophosphate
0 0.005 0.01 0.015
Figure C-14 90th percentile lead levels as a function of type of phosphate used (n=41)
• The majority of respondents use ortho-polyphosphate blends for corrosion con-

trol treatment of Source #1. As stated earlier, only 40 of the 42 systems that use
phosphates reported 90th percentile lead levels. However, one of those systems
reported using both zinc orthophosphate and polyphosphates for Source #1, thus
n=41 in Figure C-14.
Figure C-15—Comparison of 90th percentile lead levels for systems

practicing pH adjustment versus phosphates
• Review of the entire lead data set, as a function of either pH adjustment or phos-
phate addition (Figure C-15) shows systems using pH adjustment (>8.0–8.5)
and systems using phosphates had the lowest median 90th percentile lead level
(0.002 mg/L).
• Generally, systems using phosphates had overall lower 90th percentile lead levels
compared to systems practicing pH adjustment.
100
Percentile Distribution
pH>7.2–7.6
75
pH>7.6–8.0
50 pH>8.0–8.5
pH>8.5–9.0
25
Phosphates
0
0 0.003 0.006 0.009 0.012 0.015
90th Percentile Pb Levels, mg/L
Figure C-15 Comparison of 90th percentile lead levels for systems practicing pH adjustment versus
phosphates (n=73)
M58 book.indb 183 11/17/2010 4:26:12 PM

Figure C-16—90th percentile copper levels as a function of pH

(phosphate systems not included)
• Twenty-seven systems that provided 90th percentile copper levels indicated that
they use pH adjustment (no phosphates or silicates) for corrosion control.
• Median copper results shown in Figure C-16 followed the expected trend, whereby
systems with the highest finished water pH levels had the lowest 90th percentile
levels. Above the 65th percentile level, systems with pH in the range of greater than
7.2 to 7.6 appeared to have lower 90th percentile copper levels compared to systems
in the pH range greater than 7.6 to 8.0. More data analysis is needed to better under-
stand this observation.
100
Percenle Distribuon
pH>7.27.6
75
pH>7.68.0
50 pH>8.08.5
pH>8.59.0
25
pH>9.09.5
0
0 0.2 0.4 0.6 0.8 1 1.2
90th Percenle Cu Levels, mg/L
Figure C-16 90th percentile copper levels as a function of pH (n=27) (phosphate systems not
included)
Figure C-17—Percentile copper levels for systems using phosphates

• Forty systems indicated that they add phosphates for corrosion control.
• Many of these systems also practice pH adjustment (effectiveness of phosphates
as a function of finished water pH to be evaluated later).
• The median 90th percentile copper level for systems that add phosphates is
0.167 mg/L (Figure C-17).
100
Percenle Distribuon
75
50
25
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Figure C-17 90th percentile copper levels for systems using phosphates (n=40)
M58 book.indb 184 11/17/2010 4:26:13 PM

Appendix C 185
Figure C-18—Comparison of 90th percentile copper levels for systems

practicing pH adjustment versus phosphates
• Review of the entire copper data set as a function of either pH adjustment or phos-
phate addition shows systems using pH adjustment to greater than 8.0 had overall
lower copper levels compared to systems using phosphates for corrosion control.
100
pH>7.27.6
Percenle Distribuon
75 pH>7.68.0
pH>8.08.5
50
pH>8.59.0
25 pH>9.09.5
Phosphates
0
0 0.5 1 1.5
Figure C-18 Comparison of 90th percentile copper levels for systems practicing pH adjustment
versus phosphates (n=67)
Changes in corrosion control treatment. US utilities responding to the survey

were asked if any changes had been made to their corrosion control practices since origi-
nal implementation of the LCR and if they had been required to switch back to full moni-
toring after being on reduced monitoring. In summary:
• Approximately 30 percent of respondents have changed their corrosion control
practices since implementation of the LCR.
• Fine‐tuning of the corrosion control process was the most likely reason given for
changing corrosion control practices; however, many utilities listed more than one
reason. This indicates the importance of ongoing evaluation of corrosion control
practices in order to optimize for lead and copper control.
• Approximately 20 percent of utilities indicated that they have switched back to full
monitoring after being on reduced monitoring. While all utilities were in compli-
ance when the survey was filled out, six utilities listed that exceeding the action
level was the reason for their move from reduced to full monitoring.
Corrosion Control Effectiveness for Canadian Utilities

Overview of sampling techniques. In order to assess lead and copper levels in
Canadian systems, it is first important to understand the variability associated with sam-
pling techniques, because the sampling technique will have a significant impact on lead
and copper levels measured at the tap. Table C-4 summarizes sampling techniques identi-
fied by Canadian survey respondents, broken down by province/territory.
M58 book.indb 185 11/17/2010 4:26:14 PM

Table C-4 Summary of sample approaches by province/territory

Province or Territory
Newfoundland
Saskatchewan
and Labrador
Territories
Brunswick
Northwest
Columbia
Manitoba
Ontario
Alberta
Quebec
British
Scotia
Nova
New
Sample Approach
Water turns cold 1 1 1
8 hr stagnation 1*
30 min stagnation† 1 7
Fully flushed 1 1
More than 5 min 1 1 1
Flushed for 0–1 min 1 1
Flushed for 1–5 min 1 1 1
Does not sample for Pb/Cu 1 1
First draw 1
From WTP only, unless complaint 1
Total number different strategies 3 2 2 2 1 10 1 2 2 1
per province/territory
* Voluntary, not required.
† One utility listed “As per Ontario 399/07”, which is an incorrect reference. However, this utility’s location implies Ontario
170/03, which refers to water flushed for at least 5 min, then a 30–35 min stagnation time.
The results shown in Table C-4 reveal that there is significant variability with regard
to sampling approaches both between and within provinces/territories. This is an expected
result since there is no federal Lead and Copper Rule or required nationwide sampling
program. Rather, a federal guideline for lead is in place (based on a single sample per
year, using a 5-min flushed sample) that some provinces and territories have adopted (or
modified) as a regulated standard. Based on these results, and combined with the fact that
there are so few respondents within each sample approach category, it will be difficult to
evaluate corrosion control needs based on lead and copper levels reported in the survey.
Overview of Canadian corrosion control practices and lead and copper level.
Ten of the 26 utility respondents stated that they practiced corrosion control. Their pur-
pose for corrosion control and corrosion treatment techniques was summarized previously
in Table C-3. Seventeen respondents provided an average lead level, and 10 respondents
provided an average copper level. Most Canadian utilities are not required to, and do not
calculate 90th percentile levels, so results are not comparable to US results. Percentile dis-
tributions of the average lead and copper levels are provided in Figure C-19.
• Average lead levels (based on a variety of sampling protocols) are generally low,
although they approach the proposed federal first tier action level of 0.015 mg/L and
the US Action Level of 0.015 mg/L at approximately the 90th percentile. Although
it is not possible to extrapolate to determine what the 90th percentile levels would
be under the proposed federal guideline for corrosion control or under the US sam-
pling and calculation construct, it is safe to assume that 90th percentile levels
would be higher than the average lead levels shown in Figure C-19.
M58 book.indb 186 11/17/2010 4:26:14 PM

Appendix C 187
120
100
Percen�le Distribu�on
80
60
Lead
40
Copper
20
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Average Concentra�on, mg/L
Figure C-19 Percentile distribution of average lead (n=17) and copper (n=10) levels for Canadian
respondents
• Average copper levels are quite low across the distribution. The 100th percentile
copper level of 0.076 mg/L would need to increase by more than 16 times to equal
the US action level of 1.3 mg/L or increase by 13 times to exceed the federal guide-
line for copper of 1.0 mg/L.
• Seventeen of the 26 respondents indicated that they believe their lead and cop-
per levels are well below current provincial/territorial regulatory/aesthetic limits.
Three respondents believed this to be true for lead only, and one for copper only.
These results suggest that the majority of respondents do not believe they need to
take further action to control lead and/or copper levels.
Acknowledgments_______________________________________
This survey was funded by the AWWA Technical and Education Council. The authors
would like to thank the utilities in the United States and Canada that participated in the
survey. The project was managed by the AWWA Water Quality and Technology Division’s
Distribution System Water Quality Committee.
M58 book.indb 187 11/17/2010 4:26:15 PM

M58 book.indb 188 11/17/2010 4:26:15 PM

Index
Note: f. indicates a figure; t. indicates a table.
Aeration, in pH/DIC adjustment, 85–86 Biostability, 44–45

Aggressiveness Index, 16 and ammonia, 45
Alkalinity and assimilable organic carbon (AOC), 45
and buffering intensity, 140, 141f. biocorrosion (MIC) of copper pipe, 45
and carbonate balance, 33, 35f. and disinfectant residual, 45
and corrosivity, 37 and heterotrophic plate count, 46
and monitoring of corrosion control effects, 140 and nitrate, 45
Aluminum, 47 and phosphorus, 45, 46
leaching of, 9 and sulfate, 45, 46
American Water Works Association (AWWA) Black water, 2–3
corrosion inhibitor chemical standards, 86 Blue water, 3
See also North American Corrosion Control Needs Brasses, 63t., 64, 65f.
and Strategies Survey Bronze, 63t., 64, 65f.
Ammonia, and biostability, 45 Brown Deer, Wisconsin, and unaccounted-for water as
Anions, 39, 42 indicator of copper pipe corrosion, 49
chloride, 40 Buffer intensity, and carbonate balance, 33, 34f., 35f.
chloride-to-sulfate ratio, 41, 42f., 42t.
sulfate, 40 Cadmium
sulfide, 40 corrosion, 8
Anodes, 13–15, 15f. leaching of, 9
Asbestos–cement, 62, 63–64, 63t., 65f. Calcium carbonate
Asbestos–cement piping corrosion, 9, 22 failure of scales as protection against uniform
Assessing cause of corrosion and metals release, 53 corrosion, 15–16
and baseline water quality, 53 monitoring precipitation of, 134
black or yellow water as symptom, 55t. as protective measure for cement-based
blue water as symptom, 55t. materials, 64
increased tap copper concentrations as Calcium carbonate precipitation potential (CCPP), 16
symptom, 54 and carbonic acid, 106
increased tap lead concentrations as symptom, 54t. filtering apparatus, 107, 107f.
pinhole leaks in copper plumbing as symptom, 54t. and groundwater, 107, 106f.
pink water as symptom, 55t. prediction by models, 143–144
red water as symptom, 54t. test procedure, 105–107
Assessment tools, 103–104, 104t., 128 Carbon dioxide, and copper corrosion, 24
See also Bench testing; Corrosion indexes; Coupon Carbonate balance, 32, 55–56
studies; Electrochemistry (EC) monitoring; and alkalinity, 33, 35f.
Pipe loops: Pipe scale examination; Premise and buffer intensity, 33, 34f., 35f.
plumbing profiles; PRS monitoring stations; defined, 32
Reservoir profiles and dissolved inorganic carbon, 35–36, 35f., 36f.
Assimilable organic carbon (AOC), 24 and hardness, 36
and biostability, 45 and ionic strength, 34–35
and pH value, 33, 34f., 35f.
Barium, leaching of, 9 and total dissolved solids, 33–34
Bench testing, 105 Cast iron, 62, 63t., 66–67, 67f.
calcium carbonate precipitation potential test, Cathodes, 13–15, 15f.
105–107, 106f., 107f. Caustic soda (sodium/potassium hydroxide)
coagulation test for chloride-to-sulfate ratio, 108 operational aspects, 79t.
oxidation-reduction potential (ORP) test, 107 in pH/DIC adjustment, 81–82
phosphate test, 108–109 water quality and corrosion control aspects, 76t.
Biofilm, and plastic pipe, 68 CCPP. See Calcium carbonate precipitation potential
Biological regrowth, 9 Cement pipe, corrosion of, 22
189
M58 book.indb 189 11/17/2010 4:26:15 PM

Cement–mortar linings, 62, 63–64, 63t., 71, 71f., 72t. Copper corrosion, 4, 5t., 6
application of, 71f. and carbon dioxide, 24
corrosion of, 9, 22 Corrosion
Cerussite, 16–17 asbestos, 9
Chemical treatment, 74 and biological regrowth, 9
and bacteria increases, 97 cadmium, 8
chemical operational aspects, 74, 79t.–81t. of cement–mortar linings, 9, 22
and chloramine residual, 94–95 and chlorine demand, 9
and colored (red or yellow) water, 97 and color, 2–3
common corrosion control chemicals, 73t. copper, 4, 5t., 6
control of MIC, 92–93 customer and infrastructure impacts, 2–5, 9
corrosion inhibitors, 74, 86–92, 86f. defined, 2, 13
costs, comparative, 92, 93t. and distribution system deterioration, 4
costs, relative information on, 92, 93t. economic impacts, 5
DIC adjustment, 74 as electrochemical interaction between metal
and free chlorine residual, 94 surface and water, 13–15, 15f.
impact of treatment changes on corrosion control, and health concerns, 2, 9
95, 96t. and home plumbing failures, 4
passivation, 74 internal, 2
pH adjustment, 74, 81–86 iron, 8
reoptimization of, 95, 96t. lead, 4–5, 6–8, 7f.
secondary effects of, 95–97, 98t.–99t. and reequilibration of scales, 4–5, 9
treatment chemical water quality and corrosion regulatory impacts, 6–9
control aspects, 74, 75t.–78t. and taste and odor, 3
typical dry chemical feed system, 74, 75f. and water quality deterioration, 2–3, 9
typical liquid chemical feed system, 74, 74f. zinc, 8
Chloramine, and corrosion control, 19, 94–95 Corrosion by-products, 17
Chloride, 40 Corrosion control, 61–62, 100–101
and nonuniform corrosion, 24 balancing with water quality goals, 61
Chloride-to-sulfate ratio, 41, 42f., 42t. by barrier, 15, 16f.
coagulation test for, 108 coatings and linings, 62
Chlorine common chemicals in, 73t.
demand, 9 and continuing revisions, 2
and iron pipe, 66 by elimination of electron acceptors, 15, 15f.
See also Free chlorine failure of calcium carbonate scales as protection
Chromium, leaching of, 9 against uniform corrosion, 15–16
Coatings and linings, 62, 71, 72t. inhibitors, 62
cement–mortar, 71, 72t. North American needs and strategies (AWWA
coal-tar enamel, 71, 72t. survey, 2008), 173–187
epoxy, 71, 72f., 72t., 73 as ongoing, changeable task, 61
of iron pipe, 71, 71f. scaling, 4
polyethylene, 71, 72t. steps in implementation, 1, 1t., 61
Color, 2–3, 48 technique selection criteria, 100t.
Columbus (Ohio) Division of Water, and effect of and wastewater treatment, 100
ferric chloride on lead concentration and water quality modifications, 62
corrosion, 41 See also Distribution system design considerations;
Concentration cells, 23 Scales and films
Concrete, 63–64, 63t. Corrosion indexes, 105. See also Calcium Carbonate
Copper, 63t., 64, 65f. Precipitation Potential; Langelier Index
dissolved inorganic carbon in control of, 19–20 Corrosion inhibitors, 43
free chlorine in control of, 21 anodic coaters, 86f., 89–92
maximum contaminant level goal, 6 AWWA standards for chemicals, 86
orthophosphate in control of, 20 cathodic inhibitors, 86f., 87–92
polyphosphate–orthophosphate blends in chemicals, 86–92, 86f.
control of, 20 factors in success of, 87
polyphosphates in control of, 20 NSF International approval of chemicals, 86
sodium silicate in control of, 20 orthophosphate, 43, 89–91, 90f.
Copper alloys, 21 polyphosphate–orthophosphate blend, 88–89, 90f.
polyphosphates, 43–44, 87–88, 88f., 90f.
M58 book.indb 190 11/17/2010 4:26:15 PM

INDEX 191
sequestering agents, 86f. Flushing, 70

silicates, 44, 91–92 Free chlorine
sodium hexametaphosphate, 44 in corrosion control, 19, 21
stannous chloride, 92 and increase or decrease in corrosion, 94
zinc, 44 residual, 38
zinc orthophosphate, 91
Corrosivity, 36–37 Galvanic corrosion, 23
and alkalinity, 37 Galvanized iron, 63t.
defined, 36 Galvanized iron pipe, 22
and dissolved inorganic carbon, 37–38 Galvanized steel, 63t., 66, 66f.
and ion exchange softening, 37–38
and pH, 37 Hardness, and carbonate balance, 36
Coupon studies, 115–117, 117f. Heterotrophic plate count (HPC), 25, 26
coupon protocols for distribution system corrosion and biostability, 46
measures, 116, 118t. levels from tap and hydrant pilot test samples,
typical apparatus, 116, 117f. 136–137, 136f.
High density polyethylene (HDP), 62
DC Water. See Washington (D.C.) Water Hydrated lime
Disinfectant residual, and biostability, 45 operational aspects, 79t.
Dissolved inorganic carbon (DIC), 16–17, 19–20, 21 water quality and corrosion control aspects, 75t.
and carbonate balance, 35–36, 35f., 36f. Hydraulic factors, 48–49
and corrosivity, 37–38 Hydrocerussite, 16–17
Dissolved inorganic carbon (DIC) adjustment, 81
by aeration, 85–86 Infrared spectroscopy, 16
by caustic soda, 81–82 Internal Corrosion of Water Distribution Systems, 16
by lime, 82–85, 83f., 84f. Ion exchange softening, and corrosivity, 37–38
by potash, 82 Ionic strength, and carbonate balance, 34–35
by soda ash, 82 Iron, 47, 48f.
by sodium bicarbonate, 82 corrosion, 8
Dissolved oxygen, 38 dissolved inorganic carbon in control of, 21
Distribution system design considerations, 62 free chlorine in control of, 21
branched systems, 68f., 69 orthophosphate in control of, 21
engineering considerations, 69–70 polyphosphate–orthophosphate blends in
galvanic couples, 69–70, 70t. control of, 21
grid/loop systems, 68f., 69 polyphosphates in control of, 21
insulators, 69, 70 See also Cast iron; Ductile iron
pipe coatings and linings, 71–73, 71f., 72f., 72t.
system and pipe materials, 62–68, 63t., Langelier Index, 16
65f., 66f., 67f. Lead, 63t., 67f., 68
system maintenance, 70 chloramine in control of, 19
Ductile iron, 62, 63t. dissolved inorganic carbon in control of, 16–17
free chlorine in control of, 19
Electrical grounding, 24 maximum contaminant level goal, 6
Electrochemistry (EC) monitoring, 120–124 orthophosphate in control of, 17
analysis software, 122, 122t. polyphosphate–orthophosphate blends in
configuration of coupons and polarization control of, 17–18
cells, 123, 124f. polyphosphates in control of, 17–18
corrosion data from lead pipe with lead scale, and reequilibration of scales (D.C. incident), 4–5
124, 125f. sodium silicate in control of, 19
methodologies, 120, 121t. tap-water lead concentrations and LCR, 6, 7f.
pipe loop setup, 122, 124f. zinc orthophosphate in control of, 18
Pol Res tests, 123 Lead and Copper Rule (LCR), 6
polarization cell design, 124, 125f. DC WATER compliance (chronology),
Tafel tests, 123 159–160, 160f.
Erosion corrosion, 24, 25f. revisions (2007), 7–8
and tap-water lead concentrations, 6, 7f.
Ferric iron scales, 4 Lime
Ferrous iron (Fe(II)), 2–3 ball mill slakers, 83, 84f., 85
scales, 4 limestone contactors, 83
M58 book.indb 191 11/17/2010 4:26:15 PM

paste slakers, 83, 84f., 85 Naperville, Ill., and pipe requirements, 62

in pH/DIC adjustment, 82–85 Natural organic matter (NOM) removal, and
slurry detention slakers, 83, 83f., 84–85 nonuniform corrosion, 24
Lime and CO2 Nitrate, and biostability, 45
operational aspects, 79t. Nonuniform corrosion, 14t.
water quality and corrosion control aspects, 76t. and carbon dioxide, 24
Lime and soda ash and chloride, 24
operational aspects, 79t. and concentration cells, 23
water quality and corrosion control aspects, 76t. and electrical grounding, 24
Limestone filter erosion corrosion, 24, 25f.
operational aspects, 80t. galvanic, 23
water quality and corrosion control aspects, 76t. mechanisms of metal release by, 23–27
microbially influenced corrosion (MIC), 24–27, 26f.
Madison Water Utility, and increased lead levels after and natural organic matter (NOM) removal, 24
lead service line replacement, 47 Manganese, 47 and pipe installation practices, 23
Metal release, 13, 28 and pipe manufacturing quality, 23
adsorption and release by chemical scales, 27 and stray electrical currents, 24
factors in, 14t. and sulfate, 24
mechanisms in nonuniform corrosion, 23–27 North American Corrosion Control Needs and
mechanisms in uniform corrosion, 13–22 Strategies Survey, 173
and oversights in system maintenance, 27 changes in corrosion control treatment (US), 185
Metals, 46–47 corrosion control effectiveness (US), 180–185
aluminum, 47 corrosion control effectiveness and needs
dissolved, 47 (Canada), 185–187
iron, 47, 48f. corrosion control objectives, 177–179, 178f., 179t.
manganese, 47 corrosion control practices, 179, 180f.
total, 47 corrosion control practices and lead and copper
Microbially influenced corrosion (MIC), 4, 24–25, 27 levels (Canada), 186–187, 187f.
and bacteria, 25 EPA regions of respondents and SDWISFED
and biostability, 25–26 database presence, 174, 175f.
chemical treatment of, 92–93 geographic breakdown of Canadian respondents,
copper pipe pitted by, 26f., 27 174, 175f.
flushing in control of, 70 geographic breakdown of US respondents, 174,
and heterotrophic plate count (HTC), 25, 26 174f., 175f.
signs of, 25 90th percentile copper levels (pH adjustment vs.
Mild steel, 63t. phosphates in treatment), 185, 185f.
Minnesota Department of Health, and chemical 90th percentile copper levels as function of pH
inhibitors, 86 (nonphosphate systems), 184, 184f.
Monitoring corrosion control effects 90th percentile copper levels for systems using
action triggers (exceeding OCCTWQPs), 144–145, phosphates, 184, 184f.
145t., 146f., 147f. 90th percentile lead and copper levels, 180–181,
and alkalinity, 140 180f, 181f.
initial plan for, 140–141 90th percentile lead levels (pH adjustment vs.
minimum alkalinity and buffering intensity, phosphates in treatment), 183, 183f.
140, 141f. 90th percentile lead levels as function of pH (non-
and optimal water quality parameters, 140 phosphate systems), 181, 182f.
and pH, 140 90th percentile lead levels for systems using
pipe loop sampling, 143 phosphates, 182–183, 182f., 183f.
pipe scale analysis, 143 population breakdown, 175, 176f.
premise profiling, 141–143, 142f. response, 174, 174f.
PRS monitoring stations, 143 sampling techniques (Canada), 185–186, 186t.
routine plan for, 144 source water quality ranges, 175–176, 176f., 177f.
sampling techniques, 141–143
and water quality models, 143–144 Optimal corrosion control treatment water quality
and water quality target levels, 144 parameters (OCCTWQPs), 144–145
See also Electrochemistry (EC) monitoring; PRS flowchart of action taken when nonregulated
monitoring stations; Water quality monitoring parameter is exceeded, 145, 147f.
programs flowchart of action taken when regulated
parameter is exceeded, 145, 146f.
M58 book.indb 192 11/17/2010 4:26:15 PM

INDEX 193
target and excursion levels, 145, 145t. HPC levels from tap and hydrant samples,
Orthophosphate, 17, 20, 21, 89–91, 43, 89–91, 90f. 136–137, 136f.
as corrosion inhibitor for lead control, 68 maintaining chemical stability at entry point,
feed system, 90f. 133, 133f.
operational aspects, 81t. monitoring distribution system, 133–134
reasons for using, 90f. monitoring precipitation, 134, 135f.
water quality and corrosion control aspects, 76t. planning, 132
See also Polyphosphate–orthophosphate blends record keeping, 138
Ottawa, Ont., and nitrite levels causing residential mini-profiles, 137
pH depression, 140 sampling from fire hydrants, 135–137
Oxidants sampling from pipe loops, 138
dissolved oxygen, 38 sampling from reservoirs or clearwells, 134, 135f.
free chlorine residual, 38 sampling from residential taps, 137
oxidation-reduction (redox) potential, 38–39, 40f. selecting and preparing sampling sites, 134–138
Oxidation-reduction (redox) potential (ORP), 3, setting water quality target levels, 138, 139t.
38–39, 40f. use of bench test and pipe-loop data, 132
as cause of DC Water lead leaching, 162 Pipe loops, 110–111
test, 107 circulation loops, 111–113, 111f., 112f.
data from flow-through loop showing changes
pH in lead release with alternations between
and buffer intensity, 33, 34f. chlorine and chloramine, 114, 115f.
and carbonate balance, 33, 34f., 35f. flow-through, 114, 114f., 115f.
and corrosivity, 37 in evaluation of DC Water lead leaching,
increased total coliforms after change from pH 161–162, 161f.
adjustment to orthophosphate treatment, lead release during pipe-loop conditioning from
133, 134f. pipe containing lead-oxide scales, 112f., 113
and lead concentration in water, 17 Pipe scale examination, 109, 109f., 110f.
and monitoring of corrosion control effects, 140 Plastic, 63t., 67f., 68
and relationship between alkalinity and DIC, See also High density polyethylene; Polyvinyl
35, 35f. chloride
pH adjustment, 81 Polyphosphate–orthophosphate blends, 17–18, 21,
by aeration, 85–86 88–89, 90f.
by caustic soda, 81–82 as corrosion inhibitors for lead control, 68
and iron and manganese solubility, 81, 82f. reasons for using, 90f.
by lime, 82–85, 83f., 84f. Polyphosphates, 17–18, 20, 21, 43–44, 87–88, 90f.
by potash, 82 potential of secondary impacts, 99t.
potential of secondary impacts, 98t. operational aspects, 81t.
by soda ash, 82 reasons for using, 90f.
by sodium bicarbonate, 82 structures, 87, 87f.
pH stability, 149 water quality and corrosion control aspects, 77t.
and alkalinity, 149 Polyvinyl chloride (PVC), 62, 68
daily sampling for, 149 Portland (Maine) Water District, and DIC impact on
operations, 149–155 buffer intensity, 37
pump calibrations, 150f., 151 Potash, in pH/DIC adjustment, 82
requirements for, 149, 150t. Potassium hydroxide. See Caustic soda
and SCADA system in control of, 149–150 Potential-pH diagram (DC Water lead release),
startup operations, 149–150 162, 162f.
water quality parameter operations chart, 151–155, Pourbaix diagrams, 39, 40f.
152f., 153t., 154f. Premise plumbing profiles, 126–127, 127f., 128f.,
Phosphate blends, 77t. 141–143, 142f.
Phosphate test, 108–109 in evaluation of DC Water lead leaching, 161
Phosphate-based inhibitors, 66 PRS monitoring stations, 117–120, 119f., 143
Phosphate–silica mixtures, 78t. metal plates in, 119–120, 119f.
Phosphates
monitoring precipitation of, 134, 135f. Quicklime
potential of secondary impacts, 98t. operational aspects, 79t.
Phosphorus, and biostability, 45, 46 water quality and corrosion control aspects, 75t.
Pilot testing, 131
emergency response plan, 138 Red water, 2–3, 3f.
M58 book.indb 193 11/17/2010 4:26:15 PM

Reservoir profiles, 127–128 galvanized iron, 63t.

Revised Guidance Manual for Selecting Lead and galvanized steel, 63t., 66, 66f.
Copper Control Strategies (USEPA), 74 high density polyethylene (HDP), 62
Rothberg, Tamburini, and Windsor model, 143 lead, 63t., 67f., 68
Ryznar Index, 16 mild steel, 63t.
plastic, 63t., 67f., 68
Scales and films polyvinyl chloride (PVC), 62, 68
analysis by XRD and SEM, 109, 109f., 110f.
calcium carbonate, failure of, 15–16 Taste and odor, 3
and chloramine, 19 Ten States Standards, 74
dissolved inorganic carbon and other ions, 16–17, Total dissolved solids, and carbonate balance, 33–34
19–20, 21 Trace metals, 2
ferric iron, 4 Turbidity, 48
ferrous iron, 4
and free chlorine, 19, 21 Unaccounted-for water, 49
orthophosphate, 17, 20, 21 Uniform corrosion, 13, 14t.
polyphosphate–orthophosphate blends, 17–18, and cement or cement-lined pipe, 22
20, 21 and copper, 19–21
polyphosphates, 17–18, 20, 21 and copper alloys, 21
reequilibration of (D.C. incident), 4–5 failure of calcium carbonate scales as protection
sodium silicate, 19, 20 against, 15–16
zinc orthophosphate, 18 and galvanized iron pipe, 22
Scanning electron microscopy (SEM), 16 and iron, 21
in analysis of pipe wall scales, 109, 109f., 110f. and lead, 16–19
Silicate corrosion inhibitors, 67 mechanisms of metal release by, 13–22
Silicates, 44, 91–92 need for protective barrier, 13–15
operational aspects, 80t. US Army Corps of Engineers, Washington Aqueduct
potential of secondary impacts, 99t. Division, 158, 158f.
water quality and corrosion control aspects, 78t.
Soda ash Washington (D.C.) Water (DC Water)
operational aspects, 79t. corrosion evaluation tools to determine cause of
in pH/DIC adjustment, 82 lead leaching, 160–162, 161f., 162f.
water quality and corrosion control aspects, 75t. distribution system water quality target levels,
Sodium bicarbonate 171, 171t.
operational aspects, 79t. and effect of chloramine on lead dissolution, 39
in pH/DIC adjustment, 82 and electrochemistry (EC) monitoring, 120, 122,
water quality and corrosion control aspects, 75t. 124, 124f., 125f.
Sodium hexametaphosphate, 44 high ORP levels from elevated chlorine levels,
Sodium hydroxide. See Caustic soda 142f., 143
Sodium silicate, 19, 20 increased total coliforms after change from
Stafford County (Virginia) Department of Utilities pH adjustment to orthophosphate treatment,
and effect of ferric chloride on chloride-to-sulfate 133, 134f.
ratio, 41 Lead and Copper Rule compliance (chronology),
and effect of orthophosphate–polyphosphate blend 159–160, 160f.
and ferric chloride on lead concentration, 41 lead release and control case study, 157–171
Stannous chloride, 92 lead release during water hammer process in home
Stray electrical currents, 24 profiles, 142f., 143
Sulfate, 40 lead treatment test sampling and response plan,
and biostability, 45, 46 164, 167f.–170f.
and nonuniform corrosion, 24 monitoring programs, 164–171, 171t.
Sulfide, 40 orthophosphate in treatment of lead corrosion,
System and pipe materials 163–171, 163f.
asbestos–cement, 62, 63–64, 63t., 65f. partial system test of lead control,164, 165f.
bronze and brasses, 63t., 64, 65f. phosphoric feed equipment, 164, 165f.
cast iron, 62, 63t., 66–67, 67f. potential-pH diagram (lead release), 162, 162f.
cement–mortar linings, 62, 63–64, 63t. public outreach and fact sheet re lead release and
concrete, 63–64, 63t. treatment, 164, 166f.
copper, 63t., 64, 65f. reequilibration of scales and elevated-lead
ductile-iron, 62, 63t. incident, 4–5
M58 book.indb 194 11/17/2010 4:26:16 PM

INDEX 195
system description and characteristics, 158–159, example program, 52

158f., 159t. monitoring frequency, 50–51
and Technical Expert Working Group (TEWG), monitoring locations, 51–52
159–160 suggested monitoring parameters, 49–50, 50t.
and US Army Corps of Engineers, Washington Water quality parameter (WQP) operations chart,
Aqueduct Division, 158, 158f. 151–155, 152f., 153t., 154f.
Water Distribution System Monitoring, 120 WATERPRO model, 143
Water quality, 31, 55–56
additional reading, 56–57 X-ray diffraction (XRD), 16
anions, 39–42, 42f., 42t. in analysis of pipe wall scales, 109, 109f., 110f.
assessment of corrosion-related problems, 32t. in evaluation of DC Water lead leaching, 161
biostability, 44–46
carbonate balance, 32–36, 34f., 35f., 36f. Yellow water, 2–3
color, 48
corrosion inhibitors, 43–44 Zinc, 44
corrosivity, 36–38 compounds to prevent softening of cement
customer complaints, 49 pipes, 64
deterioration, and corrosion, 2–3, 9 corrosion by-products, 22
impacts on corrosion, 32t. Zinc corrosion, 8
and metals, 46–47 Zinc orthophosphate, 18, 90f., 91
and natural organic matter (NOM), 46 as corrosion inhibitor for lead control, 68
oxidants, 38–39 reasons for using, 90f.
turbidity, 48 water quality and corrosion control aspects, 77t.
Water quality monitoring programs, 32t., 56
and baseline water quality, 49
M58 book.indb 195 11/17/2010 4:26:16 PM

M58 book.indb 196 11/17/2010 4:26:16 PM

AWWA Manuals
M1, Principles of Water Rates, Fees, and Charges, M27, External Corrosion—Introduction to
Fifth Edition, 2000, #30001PA Chemistry and Control, Second Edition, 2004,
M2, Instrumentation and Control, Third Edition, #30027PA
2001, #30002PA M28, Rehabilitation of Water Mains, Second
M3, Safety Practices for Water Utilities, Sixth Edition, 2001, #30028PA
Edition, 2002, #30003PA M29, Fundamentals of Water Utility Capital
M4, Water Fluoridation Principles and Practices, Financing, Third Edition, 2008, #30029PA
Fifth Edition, 2004, #30004PA M30, Precoat Filtration, Second Edition, 1995,
M5, Water Utility Management, Second Edition, #30030PA
2006, #30005PA M31, Distribution System Requirements for Fire
M6, Water Meters—Selection, Installation, Protection, Fourth Edition, 2008, #30031PA
Testing, and Maintenance, Fourth Edition, M32, Computer Modeling of Water Distribution
1999, #30006PA Systems, Second Edition, 2005, #30032PA
M7, Problem Organisms in Water: Identification M33, Flowmeters in Water Supply, Second Edition,
and Treatment, Third Edition, 2004, #30007PA 2006, #30033PA
M9, Concrete Pressure Pipe, Third Edition, 2008, M36, Water Audits and Loss Control Programs,
#30009PA Third Edition, 2009, #30036PA
M11, Steel Pipe—A Guide for Design and M37, Operational Control of Coagulation and
Installation, Fifth Edition, 2004, #30011PA Filtration Processes, Third Edition, 2010,
M12, Simplified Procedures for Water #30037PA
Examination, Fifth Edition, 2002, #30012PA M38, Electrodialysis and Electrodialysis Reversal,
M14, Recommended Practice for Backflow First Edition, 1995, #30038PA
Prevention and Cross-Connection Control, M41, Ductile-Iron Pipe and Fittings, Third Edition,
Third Edition, 2003, #30014PA 2009, #30041PA
M17, Installation, Field Testing, and Maintenance M42, Steel Water-Storage Tanks, First Edition,
of Fire Hydrants, Fourth Edition, 2006, 1998, #30042PA
#30017PA M44, Distribution Valves: Selection, Installation,
M19, Emergency Planning for Water Utilities, Field Testing, and Maintenance, Second
Fourth Edition, 2001, #30019PA Edition, 2006, #30044PA
M20, Water Chlorination/Chloramination M45, Fiberglass Pipe Design, Second Edition, 2005,
Practices and Principles, Second Edition, #30045PA
2006, #30020PA M46, Reverse Osmosis and Nanofiltration, Second
M21, Groundwater, Third Edition, 2003, #30021PA Edition, 2007, #30046PA
M22, Sizing Water Service Lines and Meters, M47, Capital Project Delivery, Second Edition,
Second Edition, 2004, #30022PA 2010, #30047PA
M23, PVC Pipe—Design and Installation, Second M48, Waterborne Pathogens, Second Edition, 2006,
Edition, 2003, #30023PA #30048PA
M24, Planning for the Distribution of Reclaimed M49, Butterfly Valves: Torque, Head Loss, and
Water, Third Edition, 2009, #30024PA Cavitation Analysis, First Edition, 2001,
M25, Flexible-Membrane Covers and Linings for #30049PA
Potable-Water Reservoirs, Third Edition, 2000, M50, Water Resources Planning, Second Edition,
#30025PA 2007, #30050PA
To order any of these manuals or other AWWA publications, call the Bookstore toll-free at 1.800.926.7337.
197
M58 book.indb 197 11/17/2010 4:26:16 PM

M51, Air-Release, Air/Vacuum, and Combination M56, Fundamentals and Control of Nitrification
Air Valves, First Edition, 2001, #30051PA in Chloraminated Drinking Water
M52, Water Conservation Programs—A Planning Distribution Systems, First Edition, 2006,
Manual, First Edition, 2006, #30052PA #30056PA
M53, Microfiltration and Ultrafiltration M57, Algae—Source to Treatment, First Edition,
Membranes for Drinking Water, First Edition, 2010, #30057PA
2005, #30053PA M58, Internal Corrosion Control in Water
M54, Developing Rates for Small Systems, First Distribution Systems, First Edition, 2010,
Edition, 2004, #30054PA #30058PA
M55, PE Pipe—Design and Installation, First
Edition, 2006, #30055PA
To order any of these manuals or other AWWA publications, call the Bookstore toll-free at 1.800.926.7337.
M58 book.indb 198 11/17/2010 4:26:16 PM

E9sht I C C I W D S

Uploaded by

Copyright:

Available Formats

E9sht I C C I W D S

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

E9sht I C C I W D S

Uploaded by

Copyright:

Available Formats

Internal Corrosion Control in Water Distribution Systems

Improve water quality at customer taps

1P-5C-30058-12/10-SB The Authoritative Resource on Safe Water ®

M58 cover1.indd 1 11/17/2010 4:38:11 PM

AWWA MANUAL M58

M58 book.indb 1 11/17/2010 4:24:29 PM

Internal Corrosion Control in Water Distribution Systems

Copyright © 2011 American Water Works Association

Project Manager: Melissa Valentine

Library of Congress Cataloging-in-Publication Data.

Printed in the United States of America

M58 book.indb 2 11/17/2010 4:24:29 PM

List of Tables, ix

Chapter 1 Overview of Internal Corrosion Impacts

Chapter 2 Fundamentals of Internal Corrosion and Metal Release . . . . . . . . . 13

Chapter 3 Water Quality Monitoring and Assessment of Internal

Chapter 4 Corrosion Control Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

M58 book.indb 3 11/17/2010 4:24:29 PM

Chapter 6 Conducting Pilot Studies and Monitoring

Appendix A Achieving pH Stability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

Appendix B A Case Study: Causes of and Actions Taken

Appendix C North American Corrosion Control Needs and Strategies . . . . . 173

M58 book.indb 4 11/17/2010 4:24:29 PM

1-1 Red water sample from a US distribution system, 3

2-1 Typical battery, 15

3-1 Variation of buffer intensity with pH, 34

4-1 Asbestos–cement pipe, 65

M58 book.indb 5 11/17/2010 4:24:29 PM

6-1 Example of maintaining chemical (pH) stability at the entry point, 133

A-1 Example of a dual-speed pump calibration curve, 150

M58 book.indb 6 11/17/2010 4:24:29 PM

C-1 Geographic location of US participants by USEPA region, 174

M58 book.indb 7 11/17/2010 4:24:29 PM

M58 book.indb 8 11/17/2010 4:24:30 PM

1-1 Eight steps to implementing an effective corrosion control program, 1

2-1 Chapter 2 key points, 14

3-1 Chapter 3 key points, 32

4-1 Chapter 4 key points, 62

5-1 Chapter 5 key points, 104

6-1 Chapter 6 key points, 132

A-1 Summary of pH stability requirements, 150

B-1 Key characteristics of the D.C. distribution system, 159

C-1 Breakdown of survey responses, 174

M58 book.indb 9 11/17/2010 4:24:30 PM

M58 book.indb 10 11/17/2010 4:24:30 PM

M58 book.indb 11 11/17/2010 4:24:30 PM

M58 book.indb 12 11/17/2010 4:24:30 PM

A. Cantor, Process Research Solutions LLC, Madison, Wisc.

American Water Works Association, Denver, Colo.

Y. K. Cohen, Southwest Water Company, Covina, Calif.

G. Boyd – HDR Engineering, Bellevue, Wash.

G. Burlingame, Philadelphia Water Department, Philadelphia, Pa.

M58 book.indb 13 11/17/2010 4:24:30 PM

M58 book.indb 14 11/17/2010 4:24:30 PM

Table 1-1 Eight steps to implementing an effective corrosion control program

M58 book.indb 1 11/17/2010 4:24:30 PM

Selection of an effective corrosion control strategy is not a one-time event. Changes in

Purpose of This Manual__________________________________

Customer and Infrastructure Impacts of Internal

1-1 Eight steps to implementing an effective corrosion control program, 1

2-1 Chapter 2 key points, 14

3-1 Chapter 3 key points, 32

4-1 Chapter 4 key points, 62

5-1 Chapter 5 key points, 104

6-1 Chapter 6 key points, 132

A-1 Summary of pH stability requirements, 150

B-1 Key characteristics of the D.C. distribution system, 159

C-1 Breakdown of survey responses, 174